Compare commits

..

113 Commits

Author SHA1 Message Date
Konstantin Knizhnik
7f67f65d92 Use access counter for file cache pages 2022-11-04 12:55:21 +02:00
Konstantin Knizhnik
81527300ef Use access counter for giel cache pages 2022-11-04 12:22:47 +02:00
Konstantin Knizhnik
ba46de96eb Cache reconstructed pages on disk 2022-11-03 19:57:27 +02:00
Konstantin Knizhnik
e8dec662e6 Cache reconstructed pages on disk 2022-11-03 19:01:03 +02:00
Konstantin Knizhnik
ebf1972ea4 Make clippy happy 2022-10-31 23:00:54 +03:00
Joonas Koivunen
35890bb293 suggested refactoring to err out awaiting 2022-10-31 22:32:37 +03:00
Konstantin Knizhnik
8769fef1a5 Fix style 2022-10-31 19:59:34 +03:00
Konstantin Knizhnik
7452f91d5a Replace Arc<Bytes> with BYtes because Bytes maintains its own reference counter 2022-10-31 19:43:38 +03:00
Konstantin Knizhnik
6a50e1f76a Add image layer cache implementation 2022-10-29 14:29:59 +03:00
Konstantin Knizhnik
0c78aa7589 Implement page image cache with invaidtion mechanism 2022-10-28 22:06:54 +03:00
Arseny Sher
b42bf9265a Enable etcd compaction in neon_local. 2022-10-27 10:47:08 +03:00
Stas Kelvich
1f08ba5790 Avoid debian-testing packages in compute Dockerfiles
plv8 can only be built with a fairly new gold linker version. We used to install
it via binutils packages from testing, but it also updates libc and that causes
troubles in the resulting image as different extensions were built against
different libc versions. We could either use libc from debian-testing everywhere
or restrain from using testing packages and install necessary programs manually.
This patch uses the latter approach: gold for plv8 and cmake for h3 are
installed manually.

In a passing declare h3_postgis as a safe extension (previous omission).
2022-10-27 09:44:16 +03:00
bojanserafimov
0c54eb65fb Move pagestream api to libs/pageserver_api (#2698) 2022-10-26 17:32:31 -04:00
mikecaat
259a5f356e Add a docker-compose example file (#1943) (#2666)
Co-authored-by: Masahiro Ikeda <masahiro.ikeda.us@hco.ntt.co.jp>
2022-10-26 13:59:25 +03:00
Sergey Melnikov
a3cb8c11e0 Do not release to new staging proxies on release (#2685) 2022-10-25 23:51:23 +00:00
bojanserafimov
9fb2287f87 Add draw_timeline binary (#2688) 2022-10-25 11:25:22 -04:00
Alexander Bayandin
834ffe1bac Add data format backward compatibility tests (#2626) 2022-10-25 16:41:50 +02:00
Stas Kelvich
df18b041c0 Use apt version pinning instead of repo priorities
Higher `bullseye` priority doesn't works for packages installed
via `bullseye-updates`, e.g.:

```
libc-bin:
  Installed: 2.31-13+deb11u5
  Candidate: 2.35-3
  Version table:
     2.35-3 500
        500 http://ftp.debian.org/debian testing/main amd64 Packages
 *** 2.31-13+deb11u5 500
        500 http://deb.debian.org/debian bullseye-updates/main amd64 Packages
        100 /var/lib/dpkg/status
     2.31-13+deb11u4 990
        990 http://deb.debian.org/debian bullseye/main amd64 Packages
```

Try version pinning instead
2022-10-25 14:29:11 +03:00
Anastasia Lubennikova
39897105b2 Check postgres version and ensure that public schema exists
before running GRANT query on it
2022-10-25 09:55:24 +03:00
Stas Kelvich
2f399f08b2 Hotfix to disable grant create on public schema
`GRANT CREATE ON SCHEMA public` fails if there is no schema `public`.
Disable it in release for now and make a better fix later (it is
needed for v15 support).
2022-10-25 09:55:24 +03:00
Arseny Sher
9f49605041 Fix division by zero panic in determine_offloader. 2022-10-22 18:25:12 +03:00
Konstantin Knizhnik
7b6431cbd7 Disable wal_log_hints by default (#2598)
* Disable wal_log_hints by default

* Remove obsolete comment anbout wal_log_hints
2022-10-22 14:59:18 +03:00
Lassi Pölönen
321aeac3d4 Json logging capability (#2624)
* Support configuring the log format as json or plain.

Separately test json and plain logger. They would be competing on the
same global subscriber otherwise.

* Implement log_format for pageserver config

* Implement configurable log format for safekeeper.
2022-10-21 17:30:20 +00:00
Andrés
71ef7b6663 Remove cached_property package (#2673)
Co-authored-by: andres <andres.rodriguez@outlook.es>
2022-10-21 20:02:31 +03:00
Kirill Bulatov
5928cb33c5 Introduce timeline state (#2651)
Similar to https://github.com/neondatabase/neon/pull/2395, introduces a state field in Timeline, that's possible to subscribe to.

Adjusts

* walreceiver to not to have any connections if timeline is not Active
* remote storage sync to not to schedule uploads if timeline is Broken
* not to create timelines if a tenant/timeline is broken
* automatically switches timelines' states based on tenant state

Does not adjust timeline's gc, checkpointing and layer flush behaviour much, since it's not safe to cancel these processes abruptly and there's task_mgr::shutdown_tasks that does similar thing.
2022-10-21 15:51:48 +00:00
Sergey Melnikov
6ff2c61ae0 Refactor safekeeper s3 config and change it for new account (#2672) 2022-10-21 13:44:08 +00:00
Arseny Sher
7480a0338a Determine safekeeper for offloading WAL without etcd election API.
This API is rather pointless, as sane choice anyway requires knowledge of peers
status and leaders lifetime in any case can intersect, which is fine for us --
so manual elections are straightforward. Here, we deterministically choose among
the reasonably caught up safekeepers, shifting by timeline id to spread the
load.

A step towards custom broker https://github.com/neondatabase/neon/issues/2394
2022-10-21 15:33:27 +03:00
Sergey Melnikov
2709878b8b Deploy scram proxies into new account (#2643) 2022-10-21 14:21:22 +03:00
Kirill Bulatov
39e4bdb99e Actualize tenant and timeline API modifiers (#2661)
* Actualize tenant and timeline API modifiers
* Use anyhow::Result explicitly
2022-10-21 10:58:43 +00:00
Anastasia Lubennikova
52e75fead9 Use anyhow::Result explicitly 2022-10-21 12:47:06 +03:00
Anastasia Lubennikova
a347d2b6ac #2616 handle 'Unsupported pg_version' error properly 2022-10-21 12:47:06 +03:00
Heikki Linnakangas
fc4ea3553e test_gc_cutoff.py fixes (#2655)
* Fix bogus early exit from GC.

Commit 91411c415a added this failpoint, but the early exit was not
intentional.

* Cleanup test_gc_cutoff.py test.

- Remove the 'scale' parameter, this isn't a benchmark
- Tweak pgbench and pageserver options to create garbage faster that the
  the GC can collect away. The test used to take just under 5 minutes,
  which was uncomfortably close to the default 5 minute test timeout, and
  annoyingly even without the hard limit. These changes bring it down to
  about 1-2 minutes.
- Improve comments, fix typos
- Rename the failpoint. The old name, 'gc-before-save-metadata' implied
  that the failpoint was before the metadata update, but it was in fact
  much later in the function.
- Move the call to persist the metadata outside the lock, to avoid
  holding it for too long.

To verify that this test still covers the original bug,
https://github.com/neondatabase/neon/issues/2539, I commenting out
updating the metadata file like this:
```
diff --git a/pageserver/src/tenant/timeline.rs b/pageserver/src/tenant/timeline.rs
index 1e857a9a..f8a9f34a 100644
--- a/pageserver/src/tenant/timeline.rs
+++ b/pageserver/src/tenant/timeline.rs
@@ -1962,7 +1962,7 @@ impl Timeline {
         }
         // Persist the new GC cutoff value in the metadata file, before
         // we actually remove anything.
-        self.update_metadata_file(self.disk_consistent_lsn.load(), HashMap::new())?;
+        //self.update_metadata_file(self.disk_consistent_lsn.load(), HashMap::new())?;

         info!("GC starting");

```
It doesn't fail every time with that, but it did fail after about 5
runs.
2022-10-21 02:39:55 +03:00
Dmitry Rodionov
cca1ace651 make launch_wal_receiver infallible 2022-10-21 00:40:12 +03:00
Sergey Melnikov
30984c163c Fix race between pushing image to ECR and copying to dockerhub (#2662) 2022-10-20 23:01:01 +03:00
Konstantin Knizhnik
7404777efc Pin pages with speculative insert tuples to prevent their reconstruction because spec_token is not wal logged (#2657)
* Pin pages with speculative insert tuples to prevent their reconstruction because spec_token is not wal logged

refer ##2587

* Bump postgres versions
2022-10-20 20:06:05 +03:00
Heikki Linnakangas
eb1bdcc6cf If an FSM or VM page cannot be reconstructed, fill it with zeros.
If we cannot reconstruct an FSM or VM page, while creating image
layers, fill it with zeros instead. That should always be safe, for
the FSM and VM, in the sense that you won't lose actual user data. It
will get cleaned up by VACUUM later.

We had a bug with FSM/VM truncation, where we truncated the FSM and VM
at WAL replay to a smaller size than PostgreSQL originally did. We
thought was harmless, as the FSM and VM are not critical for
correctness and can be zeroed out or truncated without affecting user
data. However, it lead to a situation where PostgreSQL created
incremental WAL records for pages that we had already truncated away
in the pageserver, and when we tried to replay those WAL records, that
failed. That lead to a permanent error in image layer creation, and
prevented it from ever finishing. See
https://github.com/neondatabase/neon/issues/2601. With this patch,
those pages will be filled with zeros in the image layer, which allows
the image layer creation to finish.
2022-10-20 17:27:01 +03:00
Arthur Petukhovsky
f5ab9f761b Remove flaky checks in test_delete_force (#2567) 2022-10-20 17:14:32 +04:00
Kirill Bulatov
306a47c4fa Use uninit mark files during timeline init for atomic creation (#2489)
Part of https://github.com/neondatabase/neon/pull/2239

Regular, from scratch, timeline creation involves initdb to be run in a separate directory, data from this directory to be imported into pageserver and, finally, timeline-related background tasks to start.

This PR ensures we don't leave behind any directories that are not marked as temporary and that pageserver removes such directories on restart, allowing timeline creation to be retried with the same IDs, if needed.

It would be good to later rewrite the logic to use a temporary directory, similar what tenant creation does.
Yet currently it's harder than this change, so not done.
2022-10-20 14:19:17 +03:00
Kirill Bulatov
84c5f681b0 Fix test feature detection (#2659)
Follow-up of #2636 and #2654 , fixing the test detection feature.

Pageserver currently outputs features as

```
/target/debug/pageserver --version
Neon page server git:7734929a8202c8cc41596a861ffbe0b51b5f3cb9 failpoints: true, features: ["testing", "profiling"]
```
2022-10-20 13:44:03 +03:00
Kirill Bulatov
50297bef9f RFC about Tenant / Timeline guard objects (#2660)
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
2022-10-20 12:49:54 +03:00
Andrés
9211923bef Pageserver Python tests should not fail if the server is built with no testing feature (#2636)
Co-authored-by: andres <andres.rodriguez@outlook.es>
2022-10-20 10:46:57 +03:00
bojanserafimov
7734929a82 Remove stale todos (#2630) 2022-10-19 22:59:22 +00:00
Heikki Linnakangas
bc5ec43056 Fix flaky physical-size tests in test_timeline_size.py.
These two tests, test_timeline_physical_size_post_compaction and
test_timeline_physical_size_post_gc, assumed that after you have
waited for the WAL from a bulk insertion to arrive, and you run a
cycle of checkpoint and compaction, no new layer files are created.
Because if a new layer file is created while we are calculating the
incremental and non-incremental physical sizes, they might differ.

However, the tests used a very small checkpoint_distance, so even a
small amount of WAL generated in PostgreSQL could cause a new layer
file to be created. Autovacuum can kick in at any time, and do that.
That caused occasional failues in the test. I was able to reproduce it
reliably by adding a long delay between the incremental and
non-incremental size calculations:

```
--- a/pageserver/src/http/routes.rs
+++ b/pageserver/src/http/routes.rs
@@ -129,6 +129,9 @@ async fn build_timeline_info(
         }
     };
     let current_physical_size = Some(timeline.get_physical_size());
+    if include_non_incremental_physical_size {
+        std:🧵:sleep(std::time::Duration::from_millis(60000));
+    }

     let info = TimelineInfo {
         tenant_id: timeline.tenant_id,
```

To fix, disable autovacuum for the table. Autovacuum could still kick
in for other tables, e.g. catalog tables, but that seems less likely
to generate enough WAL to causea new layer file to be flushed.

If this continues to be a problem in the future, we could simply retry
the physical size call a few times, if there's a mismatch. A mismatch
could happen every once in a while, but it's very unlikely to happen
more than once or twice in a row.

Fixes https://github.com/neondatabase/neon/issues/2212
2022-10-19 23:50:21 +03:00
MMeent
b237feedab Add more redo metrics: (#2645)
- Measure size of redo WAL (new histogram), with bounds between 24B-32kB
- Add 2 more buckets at the upper end of the redo time histogram
  We often (>0.1% of several hours each day) take more than 250ms to do the
  redo round-trip to the postgres process. We need to measure these redo
  times more precisely.
2022-10-19 22:47:11 +02:00
Alexey Kondratov
4d1e48f3b9 [compute_ctl] Use postgres::config to properly escape database names (#2652)
We've got at least one user in production that cannot create a
database with a trailing space in the name.

This happens because we use `url` crate for manipulating the
DATABASE_URL, but it follows a standard that doesn't fit really
well with Postgres. For example, it trims all trailing spaces
from the path:

  > Remove any leading and trailing C0 control or space from input.
  > See: https://url.spec.whatwg.org/#url-parsing

But we used `set_path()` to set database name and it's totally valid
to have trailing spaces in the database name in Postgres.

Thus, use `postgres::config::Config` to modify database name in the
connection details.
2022-10-19 19:20:06 +02:00
Anastasia Lubennikova
7576b18b14 [compute_tools] fix GRANT CREATE ON SCHEMA public -
run the grant query in each database
2022-10-19 18:37:52 +03:00
Konstantin Knizhnik
6b49b370fc Fix build after applying PR #2558 2022-10-19 13:55:30 +03:00
Konstantin Knizhnik
91411c415a Persists latest_gc_cutoff_lsn before performing GC (#2558)
* Persists latest_gc_cutoff_lsn before performing GC

* Peform some refactoring and code deduplication

refer #2539

* Add test for persisting GC cutoff

* Fix python test style warnings

* Bump postgres version

* Reduce number of iterations in test_gc_cutoff test

* Bump postgres version

* Undo bumping postgres version
2022-10-19 12:32:03 +03:00
Kirill Bulatov
c67cf34040 Update GH Action version (#2646) 2022-10-19 11:16:36 +03:00
bojanserafimov
8fbe437768 Improve pageserver IO metrics (#2629) 2022-10-18 11:53:28 -04:00
Heikki Linnakangas
989d78aac8 Buffer the TCP incoming stream on libpq connections.
Reduces the number of syscalls needed to read the commands from the
compute.

Here's a snippet of strace output from the pageserver, when performing
a sequential scan on a table, with prefetch:

    3084934 recvfrom(47, "d", 1, 0, NULL, NULL) = 1
    3084934 recvfrom(47, "\0\0\0\37", 4, 0, NULL, NULL) = 4
    3084934 recvfrom(47, "\2\1\0\0\0\0\362\302\360\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\0\3", 27, 0, NULL, NULL) = 27
    3084934 pread64(28, "\0\0\0\1\0\0\0\0\0\0\0\253                    "..., 8192, 25190400) = 8192
    3084934 write(45, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\3A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3084934 poll([{fd=46, events=POLLIN}, {fd=48, events=POLLIN}], 2, 60000) = 1 ([{fd=46, revents=POLLIN}])
    3084934 read(46, "\0\0\0\0p\311q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3084934 sendto(47, "d\0\0 \5f\0\0\0\0p\311q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198
    3084934 recvfrom(47, "d", 1, 0, NULL, NULL) = 1
    3084934 recvfrom(47, "\0\0\0\37", 4, 0, NULL, NULL) = 4
    3084934 recvfrom(47, "\2\1\0\0\0\0\362\302\360\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\0\4", 27, 0, NULL, NULL) = 27
    3084934 pread64(28, "    \0=\0L\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0;;\0\0\0\4\4\0"..., 8192, 25198592) = 8192
    3084934 write(45, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\4A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3084934 poll([{fd=46, events=POLLIN}, {fd=48, events=POLLIN}], 2, 60000) = 1 ([{fd=46, revents=POLLIN}])
    3084934 read(46, "\0\0\0\0\260\344q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3084934 sendto(47, "d\0\0 \5f\0\0\0\0\260\344q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198
    3084934 recvfrom(47, "d", 1, 0, NULL, NULL) = 1
    3084934 recvfrom(47, "\0\0\0\37", 4, 0, NULL, NULL) = 4
    3084934 recvfrom(47, "\2\1\0\0\0\0\362\302\360\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\0\5", 27, 0, NULL, NULL) = 27
    3084934 write(45, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\5A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3084934 poll([{fd=46, events=POLLIN}, {fd=48, events=POLLIN}], 2, 60000) = 1 ([{fd=46, revents=POLLIN}])
    3084934 read(46, "\0\0\0\0\330\377q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3084934 sendto(47, "d\0\0 \5f\0\0\0\0\330\377q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198

This shows the interaction for three get_page_at_lsn requests. For
each request, the pageserver performs three recvfrom syscalls to read
the incoming request from the socket. After this patch, those recvfrom
calls are gone:

    3086123 read(47, "\0\0\0\0\360\222q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3086123 sendto(45, "d\0\0 \5f\0\0\0\0\360\222q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198
    3086123 pread64(29, "                                "..., 8192, 25182208) = 8192
    3086123 write(46, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\2A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3086123 poll([{fd=47, events=POLLIN}, {fd=49, events=POLLIN}], 2, 60000) = 1 ([{fd=47, revents=POLLIN}])
    3086123 read(47, "\0\0\0\0000\256q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3086123 sendto(45, "d\0\0 \5f\0\0\0\0000\256q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198
    3086123 pread64(29, "\0\0\0\1\0\0\0\0\0\0\0\253                    "..., 8192, 25190400) = 8192
    3086123 write(46, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\3A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3086123 poll([{fd=47, events=POLLIN}, {fd=49, events=POLLIN}], 2, 60000) = 1 ([{fd=47, revents=POLLIN}])
    3086123 read(47, "\0\0\0\0p\311q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237\362\0\0\237\362\0"..., 8192) = 8192
    3086123 sendto(45, "d\0\0 \5f\0\0\0\0p\311q\1\0\0\4\0\f\1\200\1\0 \4 \0\0\0\0\200\237"..., 8198, MSG_NOSIGNAL, NULL, 0) = 8198
    3086123 pread64(29, "    \0=\0L\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0;;\0\0\0\4\4\0"..., 8192, 25198592) = 8192
    3086123 write(46, "B\0\0\0\25\0\0\0\6\177\0\0002\276\0\0@\f\0\0\0\4A\0\0\32\355\0\0\0\0\1"..., 7010) = 7010
    3086123 poll([{fd=47, events=POLLIN}, {fd=49, events=POLLIN}], 2, 60000) = 1 ([{fd=47, revents=POLLIN}])

In this test, the compute sends a batch of prefetch requests, and they
are read from the socket in one syscall. That syscall was not captured
by the strace snippet above, but there are much fewer of them than
before.
2022-10-18 18:46:07 +03:00
Stas Kelvich
7ca72578f9 Enable plv8 again
Now with quickfix for https://github.com/plv8/plv8/issues/503
2022-10-18 18:34:27 +03:00
Heikki Linnakangas
41550ec8bf Remove unnecessary indirections of libpqwalproposer functions
In the Postgres backend, we cannot link directly with libpq (check the
pgsql-hackers arhive for all kinds of fun that ensued when we tried to
do that). Therefore, the libpq functions are used through the thin
wrapper functions in libpqwalreceiver.so, and libpqwalreceiver.so is
loaded dynamically. To hide the dynamic loading and make the calls
look like regular functions, we use macros to hide the function
pointers.

We had inherited the same indirections in libpqwalproposer, but it's
not needed since the neon extension is already a shared library that's
loaded dynamically. There's no problem calling the functions directly
there. Remove the indirections.
2022-10-18 18:25:30 +03:00
Sergey Melnikov
0cd2d91b9d Fix deploy-new job by installing sivel.toiletwater (#2641) 2022-10-18 14:44:19 +00:00
Sergey Melnikov
546e9bdbec Deploy storage into new account and migrate to management API v2 (#2619)
Deploy storage into new account
Migrate safekeeper and pageserver initialisation to management api v2
2022-10-18 15:52:15 +03:00
Heikki Linnakangas
59bc7e67e0 Use an optimized version of amplify_num.
Speeds up layer_map::search somewhat. I also opened a PR in the upstream
rust-amplify repository with these changes,
see https://github.com/rust-amplify/rust-amplify/pull/148. We can switch
back to upstream version when that's merged.
2022-10-18 15:00:10 +03:00
Heikki Linnakangas
2418e72649 Speed up layer_map::search, by remembering the "envelope" for each layer.
Lookups in the R-tree call the "envelope" function for every comparison,
and our envelope function isn't very cheap, so that overhead adds up.
Create the envelope once, when the layer is inserted into the tree, and
store it along with the layer. That uses some more memory per layer, but
that's not very significant.

Speeds up the search operation 2x
2022-10-18 15:00:10 +03:00
Heikki Linnakangas
80746b1c7a Add micro-benchmark for layer map search function
The test data was extracted from our pgbench benchmark project on the
captest environment, the one we use for the 'neon-captest-reuse' test.
2022-10-18 15:00:10 +03:00
Dmitry Rodionov
129f7c82b7 remove redundant expect_tenant_to_download_timeline 2022-10-18 11:21:48 +03:00
Anastasia Lubennikova
0ec5ddea0b GRANT CREATE ON SCHEMA public TO web_access 2022-10-17 22:42:51 +03:00
Kirill Bulatov
c4ee62d427 Bump clap and other minor dependencies (#2623) 2022-10-17 12:58:40 +03:00
Joonas Koivunen
c709354579 Add layer sizes to index_part.json (#2582)
This is the first step in verifying layer files. Next up on the road is
hashing the files and verifying the hashes.

The metadata additions do not require any migration. The idea is that
the change is backward and forward-compatible with regard to
`index_part.json` due to the softness of JSON schema and the
deserialization options in use.

New types added:

- LayerFileMetadata for tracking the file metadata
    - starting with only the file size
    - in future hopefully a sha256 as well
- IndexLayerMetadata, the serialized counterpart of LayerFileMetadata

LayerFileMetadata needing to have all fields Option is a problem but
that is not possible to handle without conflicting a lot more with other
ongoing work.

Co-authored-by: Kirill Bulatov <kirill@neon.tech>
2022-10-17 12:21:04 +03:00
Lassi Pölönen
5d6553d41d Fix pageserver configuration generation bug (#2584)
* We had an issue with `lineinfile` usage for pageserver configuration
file: if the S3 bucket related values were changed, it would have
resulted in duplicate keys, resulting in invalid toml.

So to fix the issue, we should keep the configuration in structured
format (yaml in this case) so we can always generate syntactically
correct toml.

Inventories are converted to yaml just so that it's easier to maintain
the configuration there. Another alternative would have been a separate
variable files.

* Keep the ansible collections dir, but locally installed collections
should not be tracked.
2022-10-16 11:37:10 +00:00
Kirill Bulatov
f03b7c3458 Bump regular dependencies (#2618)
* etcd-client is not updated, since we plan to replace it with another client and the new version errors with some missing prost library error
* clap had released another major update that requires changing every CLI declaration again, deserves a separate PR
2022-10-15 01:55:31 +03:00
Heikki Linnakangas
9c24de254f Add description and license fields to OpenAPI spec.
These were added earlier to the control plane's copy of this file.
This is the master version of this file, so let's keep it in sync.
2022-10-14 18:37:58 +03:00
Heikki Linnakangas
538876650a Merge 'local' and 'remote' parts of TimelineInfo into one struct.
The 'local' part was always filled in, so that was easy to merge into
into the TimelineInfo itself. 'remote' only contained two fields,
'remote_consistent_lsn' and 'awaits_download'. I made
'remote_consistent_lsn' an optional field, and 'awaits_download' is now
false if the timeline is not present remotely.

However, I kept stub versions of the 'local' and 'remote' structs for
backwards-compatibility, with a few fields that are actively used by
the control plane. They just duplicate the fields from TimelineInfo
now. They can be removed later, once the control plane has been
updated to use the new fields.
2022-10-14 18:37:14 +03:00
Heikki Linnakangas
500239176c Make TimelineInfo.local field mandatory.
It was only None when you queried the status of a timeline with
'timeline_detail' mgmt API call, and it was still being downloaded. You
can check for that status with the 'tenant_status' API call instead,
checking for has_in_progress_downloads field.

Anothere case was if an error happened while trying to get the current
logical size, in a 'timeline_detail' request. It might make sense to
tolerate such errors, and leave the fields we cannot fill in as empty,
None, 0 or similar, but it doesn't make sense to me to leave the whole
'local' struct empty in tht case.
2022-10-14 18:37:14 +03:00
Anastasia Lubennikova
ee64a6b80b Fix CI: push versioned compute images to production ECR 2022-10-14 18:12:50 +03:00
Anastasia Lubennikova
a13b486943 Bump vendor/postgres-v15. Rebase to 15.0 2022-10-14 18:12:50 +03:00
Arseny Sher
9fe4548e13 Reimplement explicit timeline creation on safekeepers.
With the ability to pass commit_lsn. This allows to perform project WAL recovery
through different (from the original) set of safekeepers (or under different
ttid) by
1) moving WAL files to s3 under proper ttid;
2) explicitly creating timeline on safekeepers, setting commit_lsn to the
latest point;
3) putting the lastest .parital file to the timeline directory on safekeepers, if
desired.

Extend test_s3_wal_replay to exersise this behaviour.

Also extends timeline_status endpoint to return postgres information.
2022-10-13 21:43:10 +04:00
Heikki Linnakangas
14c623b254 Make it possible to build with old cargo version.
I'm using the Rust compiler and cargo versions from Debian packages,
but the latest available cargo Debian package is quite old, version
1.57.  The 'named-profiles' features was not stabilized at that
version yet, so ever since commit a463749f5, I've had to manually add
this line to the Cargo.toml file to compile. I've been wishing that
someone would update the cargo Debian package, but it doesn't seem to
be happening any time soon.

This doesn't seem to bother anyone else but me, but it shouldn't hurt
anyone else either. If there was a good reason, I could install a
newer cargo version with 'rustup', but if all we need is this one line
in Cargo.toml, I'd prefer to continue using the Debian packages.
2022-10-13 15:17:00 +03:00
Alexander Bayandin
ebf54b0de0 Nightly Benchmarks: Add 50 GB projects (#2612) 2022-10-13 10:00:29 +01:00
Andrés
09dda35dac Return broken tenants due to non existing timelines dir (#2552) (#2575)
Co-authored-by: andres <andres.rodriguez@outlook.es>
2022-10-12 22:28:39 +03:00
Dmitry Ivanov
6ace79345d [proxy] Add more context to console requests logging (#2583) 2022-10-12 21:00:44 +03:00
danieltprice
771e61425e Update release-pr.md (#2600)
Update the Release Notes PR example that is referenced from the checklist. The Release Notes file structure changed recently.
2022-10-12 08:38:28 -03:00
Alexander Bayandin
93775f6ca7 GitHub Actions: replace deprecated set-output with GITHUB_OUTPUT (#2608) 2022-10-12 10:22:24 +01:00
Arseny Sher
6d0dacc4ce Recreate timeline on pageserver in s3_wal_replay test.
That's closer to real usage than switching to brand new pageserver.
2022-10-12 11:46:21 +04:00
Heikki Linnakangas
e5e40a31f4 Clean up terms "delete timeline" and "detach tenant".
You cannot attach/detach an individual timeline, attach/detach always
applies to the whole tenant. However, you can *delete* a single timeline
from a tenant. Fix some comments and error messages that confused these
two operations.
2022-10-11 17:47:41 +03:00
Heikki Linnakangas
676c63c329 Improve comments. 2022-10-11 17:47:41 +03:00
Heikki Linnakangas
47366522a8 Make the return type of 'list_timelines' simpler.
It's enough to return just the Timeline references. You can get the
timeline's ID easily from Timeline.
2022-10-11 17:47:41 +03:00
Heikki Linnakangas
db26bc49cc Remove obsolete FIXME comment.
Commit c634cb1d36 removed the trait and changed the function to return
a &TimelineWriter, as the FIXME said we should do, but forgot to remove
the FIXME.
2022-10-11 17:47:41 +03:00
Lassi Pölönen
e520293090 Add build info metric to pageserver, safekeeper and proxy (#2596)
* Test that we emit build info metric for pageserver, safekeeper and proxy with some non-zero length revision label

* Emit libmetrics_build_info on startup of pageserver, safekeeper and
proxy with label "revision" which tells the git revision.
2022-10-11 09:54:32 +03:00
Sergey Melnikov
241e549757 Switch neon-stress etcd to dedicatd instance (#2602) 2022-10-10 22:07:19 +00:00
Sergey Melnikov
34bea270f0 Fix POSTGRES_DISTRIB_DIR for benchmarks on ec2 runner (#2594) 2022-10-10 09:12:50 +00:00
Kirill Bulatov
13f0e7a5b4 Deploy pageserver_binutils to the envs 2022-10-09 08:21:11 +03:00
Kirill Bulatov
3e35f10adc Add a script to reformat the project 2022-10-09 08:21:11 +03:00
Kirill Bulatov
3be3bb7730 Be more verbose with initdb for pageserver timeline creation 2022-10-09 08:21:11 +03:00
Kirill Bulatov
01d2c52c82 Tidy up feature reporting 2022-10-09 08:21:11 +03:00
Kirill Bulatov
9f79e7edea Merge pageserver helper binaries and provide it for deployment (#2590) 2022-10-08 12:42:17 +00:00
Heikki Linnakangas
a22165d41e Add tests for comparing root and child branch performance.
Author: Thang Pham <thang@neon.tech>
2022-10-08 10:07:33 +03:00
Arseny Sher
725be60bb7 Storage messaging rfc 2. 2022-10-07 21:22:17 +04:00
Dmitry Ivanov
e516c376d6 [proxy] Improve logging (#2554)
* [proxy] Use `tracing::*` instead of `println!` for logging

* Fix a minor misnomer

* Log more stuff
2022-10-07 14:34:57 +03:00
Kirill Bulatov
8e51c27e1a Restore artifact versions (#2578)
Context: https://github.com/neondatabase/neon/pull/2128/files#r989489965

Co-authored-by: Rory de Zoete <rory@neon.tech>
2022-10-07 10:58:31 +00:00
Heikki Linnakangas
9e1eb69d55 Increase default compaction_period setting to 20 s.
The previous default of 1 s caused excessive CPU usage when there were
a lot of projects. Polling every timeline once a second was too aggressive
so let's reduce it.

Fixes https://github.com/neondatabase/neon/issues/2542, but we
probably also want do to something so that we don't poll timelines
that have received no new WAL or layers since last check.
2022-10-07 13:55:19 +03:00
Arthur Petukhovsky
687ba81366 Display sync safekeepers output in compute_ctl (#2571)
Pipe postgres output to compute_ctl stdout and create a test to check that compute_ctl works and prints postgres logs.
2022-10-06 13:53:52 +00:00
Andrés
47bae68a2e Make get_lsn_by_timestamp available in mgmt API (#2536) (#2560)
Co-authored-by: andres <andres.rodriguez@outlook.es>
2022-10-06 12:42:50 +03:00
Joonas Koivunen
e8b195acb7 fix: apply notify workaround on m1 mac docker (#2564)
workaround as discussed in the notify repository.
2022-10-06 11:13:40 +03:00
Anastasia Lubennikova
254cb7dc4f Update CI script to push compute-node-v15 to dockerhub 2022-10-06 10:50:08 +03:00
Anastasia Lubennikova
ed85d97f17 bump vendor/postgres-v15. Rebase it to Stamp 15rc2 2022-10-06 10:50:08 +03:00
Anastasia Lubennikova
4a216c5f7f Use PostGIS 3.3.1 that is compatible with pg 15 2022-10-06 10:50:08 +03:00
Anastasia Lubennikova
c5a428a61a Update Dockerfile.compute-node-v15 to match v14 version.
Fix build script to promote the image for v15 to neon dockerhub
2022-10-06 10:50:08 +03:00
Konstantin Knizhnik
ff8c481777 Normalize last_record LSN in wal receiver (#2529)
* Add test for branching on page boundary

* Normalize start recovery point

Co-authored-by: Heikki Linnakangas <heikki@neon.tech>

Co-authored-by:  Thang Pham <thang@neon.tech>
2022-10-06 09:01:56 +03:00
Arthur Petukhovsky
f25dd75be9 Fix deadlock in safekeeper metrics (#2566)
We had a problem where almost all of the threads were waiting on a futex syscall. More specifically:
- `/metrics` handler was inside `TimelineCollector::collect()`, waiting on a mutex for a single Timeline
- This exact timeline was inside `control_file::FileStorage::persist()`, waiting on a mutex for Lazy initialization of `PERSIST_CONTROL_FILE_SECONDS`
- `PERSIST_CONTROL_FILE_SECONDS: Lazy<Histogram>` was blocked on `prometheus::register`
- `prometheus::register` calls `DEFAULT_REGISTRY.write().register()` to take a write lock on Registry and add a new metric
- `DEFAULT_REGISTRY` lock was already taken inside `DEFAULT_REGISTRY.gather()`, which was called by `/metrics` handler to collect all metrics

This commit creates another Registry with a separate lock, to avoid deadlock in a case where `TimelineCollector` triggers registration of new metrics inside default registry.
2022-10-06 01:07:02 +03:00
Sergey Melnikov
b99bed510d Move proxies to neon-proxy namespace (#2555) 2022-10-05 16:14:09 +03:00
sharnoff
580584c8fc Remove control_plane deps on pageserver/safekeeper (#2513)
Creates new `pageserver_api` and `safekeeper_api` crates to serve as the
shared dependencies. Should reduce both recompile times and cold compile
times.

Decreases the size of the optimized `neon_local` binary: 380M -> 179M.
No significant changes for anything else (mostly as expected).
2022-10-04 11:14:45 -07:00
Kirill Bulatov
d823e84ed5 Allow attaching tenants with zero timelines 2022-10-04 18:13:51 +03:00
Kirill Bulatov
231dfbaed6 Do not remove empty timelines/ directory for tenants 2022-10-04 18:13:51 +03:00
Dmitry Rodionov
5cf53786f9 Improve pytest ergonomics
1. Disable perf tests by default
2. Add instruction to run tests in parallel
2022-10-04 14:53:01 +03:00
Heikki Linnakangas
9b9bbad462 Use 'notify' crate to wait for PostgreSQL startup.
Compute node startup time is very important. After launching
PostgreSQL, use 'notify' to be notified immediately when it has
updated the PID file, instead of polling. The polling loop had 100 ms
interval so this shaves up to 100 ms from the startup time.
2022-10-04 13:00:15 +03:00
Heikki Linnakangas
537b2c1ae6 Remove unnecessary check for open PostgreSQL TCP port.
The loop checked if the TCP port is open for connections, by trying to
connect to it. That seems unnecessary. By the time the postmaster.pid
file says that it's ready, the port should be open. Remove that check.
2022-10-04 12:09:13 +03:00
Joonas Koivunen
31123d1fa8 Silence clippies, minor doc fix (#2543)
* doc: remove stray backtick

* chore: clippy::let_unit_value

* chore: silence useless_transmute, duplicate_mod

* chore: remove allowing deref_nullptr

not needed since bindgen 0.60.0.

* chore: remove repeated allowed lints

they are already allowed from the crate root.
2022-10-03 17:44:17 +03:00
Kirill Bulatov
4f2ac51bdd Bump rustc to 1.61 2022-10-03 16:36:03 +03:00
Kirill Bulatov
7b2f9dc908 Reuse existing tenants during attach (#2540) 2022-10-03 13:33:55 +03:00
195 changed files with 15087 additions and 4399 deletions

View File

@@ -10,7 +10,7 @@
<!-- List everything that should be done **before** release, any issues / setting changes / etc -->
### Checklist after release
- [ ] Based on the merged commits write release notes and open a PR into `website` repo ([example](https://github.com/neondatabase/website/pull/120/files))
- [ ] Based on the merged commits write release notes and open a PR into `website` repo ([example](https://github.com/neondatabase/website/pull/219/files))
- [ ] Check [#dev-production-stream](https://neondb.slack.com/archives/C03F5SM1N02) Slack channel
- [ ] Check [stuck projects page](https://console.neon.tech/admin/projects?sort=last_active&order=desc&stuck=true)
- [ ] Check [recent operation failures](https://console.neon.tech/admin/operations?action=create_timeline%2Cstart_compute%2Cstop_compute%2Csuspend_compute%2Capply_config%2Cdelete_timeline%2Cdelete_tenant%2Ccreate_branch%2Ccheck_availability&sort=updated_at&order=desc&had_retries=some)

View File

@@ -47,7 +47,7 @@ runs:
else
key=branch-$(echo ${GITHUB_REF#refs/heads/} | tr -c "[:alnum:]._-" "-")
fi
echo "::set-output name=KEY::${key}"
echo "KEY=${key}" >> $GITHUB_OUTPUT
- uses: actions/setup-java@v3
if: ${{ inputs.action == 'generate' }}
@@ -186,7 +186,7 @@ runs:
aws s3 cp --only-show-errors ./index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"
echo "[Allure Report](${REPORT_URL})" >> ${GITHUB_STEP_SUMMARY}
echo "::set-output name=report-url::${REPORT_URL}"
echo "report-url=${REPORT_URL}" >> $GITHUB_OUTPUT
- name: Release Allure lock
if: ${{ inputs.action == 'generate' && always() }}

View File

@@ -34,7 +34,7 @@ runs:
S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${PREFIX%$GITHUB_RUN_ATTEMPT} | jq -r '.Contents[].Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
if [ -z "${S3_KEY}" ]; then
if [ "${SKIP_IF_DOES_NOT_EXIST}" = "true" ]; then
echo '::set-output name=SKIPPED::true'
echo 'SKIPPED=true' >> $GITHUB_OUTPUT
exit 0
else
echo 2>&1 "Neither s3://${BUCKET}/${PREFIX}/${FILENAME} nor its version from previous attempts exist"
@@ -42,7 +42,7 @@ runs:
fi
fi
echo '::set-output name=SKIPPED::false'
echo 'SKIPPED=false' >> $GITHUB_OUTPUT
mkdir -p $(dirname $ARCHIVE)
time aws s3 cp --only-show-errors s3://${BUCKET}/${S3_KEY} ${ARCHIVE}

View File

@@ -41,8 +41,8 @@ runs:
;;
esac
echo "::set-output name=api_host::${API_HOST}"
echo "::set-output name=region_id::${REGION_ID}"
echo "api_host=${API_HOST}" >> $GITHUB_OUTPUT
echo "region_id=${REGION_ID}" >> $GITHUB_OUTPUT
env:
ENVIRONMENT: ${{ inputs.environment }}
REGION_ID: ${{ inputs.region_id }}
@@ -72,10 +72,10 @@ runs:
dsn=$(echo $project | jq --raw-output '.roles[] | select(.name != "web_access") | .dsn')/main
echo "::add-mask::${dsn}"
echo "::set-output name=dsn::${dsn}"
echo "dsn=${dsn}" >> $GITHUB_OUTPUT
project_id=$(echo $project | jq --raw-output '.id')
echo "::set-output name=project_id::${project_id}"
echo "project_id=${project_id}" >> $GITHUB_OUTPUT
env:
API_KEY: ${{ inputs.api_key }}
API_HOST: ${{ steps.parse-input.outputs.api_host }}

View File

@@ -32,7 +32,7 @@ runs:
;;
esac
echo "::set-output name=api_host::${API_HOST}"
echo "api_host=${API_HOST}" >> $GITHUB_OUTPUT
env:
ENVIRONMENT: ${{ inputs.environment }}

View File

@@ -73,6 +73,13 @@ runs:
shell: bash -euxo pipefail {0}
run: ./scripts/pysync
- name: Download compatibility snapshot for Postgres 14
uses: ./.github/actions/download
with:
name: compatibility-snapshot-${{ inputs.build_type }}-pg14
path: /tmp/compatibility_snapshot_pg14
prefix: latest
- name: Run pytest
env:
NEON_BIN: /tmp/neon/bin
@@ -80,6 +87,8 @@ runs:
BUILD_TYPE: ${{ inputs.build_type }}
AWS_ACCESS_KEY_ID: ${{ inputs.real_s3_access_key_id }}
AWS_SECRET_ACCESS_KEY: ${{ inputs.real_s3_secret_access_key }}
COMPATIBILITY_SNAPSHOT_DIR: /tmp/compatibility_snapshot_pg14
ALLOW_BREAKING_CHANGES: contains(github.event.pull_request.labels.*.name, 'breaking changes')
shell: bash -euxo pipefail {0}
run: |
# PLATFORM will be embedded in the perf test report
@@ -154,6 +163,15 @@ runs:
scripts/generate_and_push_perf_report.sh
fi
- name: Upload compatibility snapshot for Postgres 14
if: github.ref_name == 'release'
uses: ./.github/actions/upload
with:
name: compatibility-snapshot-${{ inputs.build_type }}-pg14-${{ github.run_id }}
# The path includes a test name (test_prepare_snapshot) and directory that the test creates (compatibility_snapshot_pg14), keep the path in sync with the test
path: /tmp/test_output/test_prepare_snapshot/compatibility_snapshot_pg14/
prefix: latest
- name: Create Allure report
if: always()
uses: ./.github/actions/allure-report

View File

@@ -2,3 +2,6 @@ zenith_install.tar.gz
.zenith_current_version
neon_install.tar.gz
.neon_current_version
collections/*
!collections/.keep

View File

@@ -3,6 +3,7 @@
localhost_warning = False
host_key_checking = False
timeout = 30
collections_paths = ./collections
[ssh_connection]
ssh_args = -F ./ansible.ssh.cfg

0
.github/ansible/collections/.keep vendored Normal file
View File

View File

@@ -1,7 +1,7 @@
- name: Upload Neon binaries
hosts: storage
gather_facts: False
remote_user: admin
remote_user: "{{ remote_user }}"
tasks:
@@ -14,7 +14,8 @@
- safekeeper
- name: inform about versions
debug: msg="Version to deploy - {{ current_version }}"
debug:
msg: "Version to deploy - {{ current_version }}"
tags:
- pageserver
- safekeeper
@@ -35,7 +36,7 @@
- name: Deploy pageserver
hosts: pageservers
gather_facts: False
remote_user: admin
remote_user: "{{ remote_user }}"
tasks:
@@ -63,15 +64,29 @@
tags:
- pageserver
- name: update remote storage (s3) config
lineinfile:
path: /storage/pageserver/data/pageserver.toml
line: "{{ item }}"
loop:
- "[remote_storage]"
- "bucket_name = '{{ bucket_name }}'"
- "bucket_region = '{{ bucket_region }}'"
- "prefix_in_bucket = '{{ inventory_hostname }}'"
- name: read the existing remote pageserver config
ansible.builtin.slurp:
src: /storage/pageserver/data/pageserver.toml
register: _remote_ps_config
tags:
- pageserver
- name: parse the existing pageserver configuration
ansible.builtin.set_fact:
_existing_ps_config: "{{ _remote_ps_config['content'] | b64decode | sivel.toiletwater.from_toml }}"
tags:
- pageserver
- name: construct the final pageserver configuration dict
ansible.builtin.set_fact:
pageserver_config: "{{ pageserver_config_stub | combine({'id': _existing_ps_config.id }) }}"
tags:
- pageserver
- name: template the pageserver config
template:
src: templates/pageserver.toml.j2
dest: /storage/pageserver/data/pageserver.toml
become: true
tags:
- pageserver
@@ -109,7 +124,7 @@
- name: Deploy safekeeper
hosts: safekeepers
gather_facts: False
remote_user: admin
remote_user: "{{ remote_user }}"
tasks:

View File

@@ -23,6 +23,7 @@ docker cp ${ID}:/data/postgres_install.tar.gz .
tar -xzf postgres_install.tar.gz -C neon_install
mkdir neon_install/bin/
docker cp ${ID}:/usr/local/bin/pageserver neon_install/bin/
docker cp ${ID}:/usr/local/bin/pageserver_binutils neon_install/bin/
docker cp ${ID}:/usr/local/bin/safekeeper neon_install/bin/
docker cp ${ID}:/usr/local/bin/proxy neon_install/bin/
docker cp ${ID}:/usr/local/v14/bin/ neon_install/v14/bin/

View File

@@ -1,20 +0,0 @@
[pageservers]
neon-stress-ps-1 console_region_id=1
neon-stress-ps-2 console_region_id=1
[safekeepers]
neon-stress-sk-1 console_region_id=1
neon-stress-sk-2 console_region_id=1
neon-stress-sk-3 console_region_id=1
[storage:children]
pageservers
safekeepers
[storage:vars]
env_name = neon-stress
console_mgmt_base_url = http://neon-stress-console.local
bucket_name = neon-storage-ireland
bucket_region = eu-west-1
etcd_endpoints = etcd-stress.local:2379
safekeeper_enable_s3_offload = false

31
.github/ansible/neon-stress.hosts.yaml vendored Normal file
View File

@@ -0,0 +1,31 @@
storage:
vars:
bucket_name: neon-storage-ireland
bucket_region: eu-west-1
console_mgmt_base_url: http://neon-stress-console.local
etcd_endpoints: neon-stress-etcd.local:2379
safekeeper_enable_s3_offload: 'false'
pageserver_config_stub:
pg_distrib_dir: /usr/local
remote_storage:
bucket_name: "{{ bucket_name }}"
bucket_region: "{{ bucket_region }}"
prefix_in_bucket: "{{ inventory_hostname }}"
safekeeper_s3_prefix: neon-stress/wal
hostname_suffix: ".local"
remote_user: admin
children:
pageservers:
hosts:
neon-stress-ps-1:
console_region_id: aws-eu-west-1
neon-stress-ps-2:
console_region_id: aws-eu-west-1
safekeepers:
hosts:
neon-stress-sk-1:
console_region_id: aws-eu-west-1
neon-stress-sk-2:
console_region_id: aws-eu-west-1
neon-stress-sk-3:
console_region_id: aws-eu-west-1

View File

@@ -1,20 +0,0 @@
[pageservers]
#zenith-1-ps-1 console_region_id=1
zenith-1-ps-2 console_region_id=1
zenith-1-ps-3 console_region_id=1
[safekeepers]
zenith-1-sk-1 console_region_id=1
zenith-1-sk-2 console_region_id=1
zenith-1-sk-3 console_region_id=1
[storage:children]
pageservers
safekeepers
[storage:vars]
env_name = prod-1
console_mgmt_base_url = http://console-release.local
bucket_name = zenith-storage-oregon
bucket_region = us-west-2
etcd_endpoints = zenith-1-etcd.local:2379

33
.github/ansible/production.hosts.yaml vendored Normal file
View File

@@ -0,0 +1,33 @@
---
storage:
vars:
console_mgmt_base_url: http://console-release.local
bucket_name: zenith-storage-oregon
bucket_region: us-west-2
etcd_endpoints: zenith-1-etcd.local:2379
pageserver_config_stub:
pg_distrib_dir: /usr/local
remote_storage:
bucket_name: "{{ bucket_name }}"
bucket_region: "{{ bucket_region }}"
prefix_in_bucket: "{{ inventory_hostname }}"
safekeeper_s3_prefix: prod-1/wal
hostname_suffix: ".local"
remote_user: admin
children:
pageservers:
hosts:
zenith-1-ps-2:
console_region_id: aws-us-west-2
zenith-1-ps-3:
console_region_id: aws-us-west-2
safekeepers:
hosts:
zenith-1-sk-1:
console_region_id: aws-us-west-2
zenith-1-sk-2:
console_region_id: aws-us-west-2
zenith-1-sk-3:
console_region_id: aws-us-west-2

View File

@@ -12,18 +12,19 @@ cat <<EOF | tee /tmp/payload
"version": 1,
"host": "${HOST}",
"port": 6400,
"region_id": {{ console_region_id }},
"region_id": "{{ console_region_id }}",
"instance_id": "${INSTANCE_ID}",
"http_host": "${HOST}",
"http_port": 9898
"http_port": 9898,
"active": false
}
EOF
# check if pageserver already registered or not
if ! curl -sf -X PATCH -d '{}' {{ console_mgmt_base_url }}/api/v1/pageservers/${INSTANCE_ID} -o /dev/null; then
if ! curl -sf -H "Authorization: Bearer {{ CONSOLE_API_TOKEN }}" {{ console_mgmt_base_url }}/management/api/v2/pageservers/${INSTANCE_ID} -o /dev/null; then
# not registered, so register it now
ID=$(curl -sf -X POST {{ console_mgmt_base_url }}/api/v1/pageservers -d@/tmp/payload | jq -r '.ID')
ID=$(curl -sf -X POST -H "Authorization: Bearer {{ CONSOLE_API_TOKEN }}" {{ console_mgmt_base_url }}/management/api/v2/pageservers -d@/tmp/payload | jq -r '.id')
# init pageserver
sudo -u pageserver /usr/local/bin/pageserver -c "id=${ID}" -c "pg_distrib_dir='/usr/local'" --init -D /storage/pageserver/data

View File

@@ -14,18 +14,18 @@ cat <<EOF | tee /tmp/payload
"host": "${HOST}",
"port": 6500,
"http_port": 7676,
"region_id": {{ console_region_id }},
"region_id": "{{ console_region_id }}",
"instance_id": "${INSTANCE_ID}",
"availability_zone_id": "${AZ_ID}"
"availability_zone_id": "${AZ_ID}",
"active": false
}
EOF
# check if safekeeper already registered or not
if ! curl -sf -X PATCH -d '{}' {{ console_mgmt_base_url }}/api/v1/safekeepers/${INSTANCE_ID} -o /dev/null; then
if ! curl -sf -H "Authorization: Bearer {{ CONSOLE_API_TOKEN }}" {{ console_mgmt_base_url }}/management/api/v2/safekeepers/${INSTANCE_ID} -o /dev/null; then
# not registered, so register it now
ID=$(curl -sf -X POST {{ console_mgmt_base_url }}/api/v1/safekeepers -d@/tmp/payload | jq -r '.ID')
ID=$(curl -sf -X POST -H "Authorization: Bearer {{ CONSOLE_API_TOKEN }}" {{ console_mgmt_base_url }}/management/api/v2/safekeepers -d@/tmp/payload | jq -r '.id')
# init safekeeper
sudo -u safekeeper /usr/local/bin/safekeeper --id ${ID} --init -D /storage/safekeeper/data
fi

3
.github/ansible/ssm_config vendored Normal file
View File

@@ -0,0 +1,3 @@
ansible_connection: aws_ssm
ansible_aws_ssm_bucket_name: neon-dev-bucket
ansible_python_interpreter: /usr/bin/python3

View File

@@ -1,25 +0,0 @@
[pageservers]
#zenith-us-stage-ps-1 console_region_id=27
zenith-us-stage-ps-2 console_region_id=27
zenith-us-stage-ps-3 console_region_id=27
zenith-us-stage-ps-4 console_region_id=27
zenith-us-stage-test-ps-1 console_region_id=28
[safekeepers]
zenith-us-stage-sk-4 console_region_id=27
zenith-us-stage-sk-5 console_region_id=27
zenith-us-stage-sk-6 console_region_id=27
zenith-us-stage-test-sk-1 console_region_id=28
zenith-us-stage-test-sk-2 console_region_id=28
zenith-us-stage-test-sk-3 console_region_id=28
[storage:children]
pageservers
safekeepers
[storage:vars]
env_name = us-stage
console_mgmt_base_url = http://console-staging.local
bucket_name = zenith-staging-storage-us-east-1
bucket_region = us-east-1
etcd_endpoints = zenith-us-stage-etcd.local:2379

34
.github/ansible/staging.hosts.yaml vendored Normal file
View File

@@ -0,0 +1,34 @@
storage:
vars:
bucket_name: zenith-staging-storage-us-east-1
bucket_region: us-east-1
console_mgmt_base_url: http://console-staging.local
etcd_endpoints: zenith-us-stage-etcd.local:2379
pageserver_config_stub:
pg_distrib_dir: /usr/local
remote_storage:
bucket_name: "{{ bucket_name }}"
bucket_region: "{{ bucket_region }}"
prefix_in_bucket: "{{ inventory_hostname }}"
safekeeper_s3_prefix: us-stage/wal
hostname_suffix: ".local"
remote_user: admin
children:
pageservers:
hosts:
zenith-us-stage-ps-2:
console_region_id: aws-us-east-1
zenith-us-stage-ps-3:
console_region_id: aws-us-east-1
zenith-us-stage-ps-4:
console_region_id: aws-us-east-1
safekeepers:
hosts:
zenith-us-stage-sk-4:
console_region_id: aws-us-east-1
zenith-us-stage-sk-5:
console_region_id: aws-us-east-1
zenith-us-stage-sk-6:
console_region_id: aws-us-east-1

View File

@@ -0,0 +1,32 @@
storage:
vars:
bucket_name: neon-staging-storage-us-east-2
bucket_region: us-east-2
console_mgmt_base_url: http://console-staging.local
etcd_endpoints: etcd-0.us-east-2.aws.neon.build:2379
pageserver_config_stub:
pg_distrib_dir: /usr/local
remote_storage:
bucket_name: "{{ bucket_name }}"
bucket_region: "{{ bucket_region }}"
prefix_in_bucket: "pageserver/v1"
safekeeper_s3_prefix: safekeeper/v1/wal
hostname_suffix: ""
remote_user: ssm-user
ansible_aws_ssm_region: us-east-2
console_region_id: aws-us-east-2
children:
pageservers:
hosts:
pageserver-0.us-east-2.aws.neon.build:
ansible_host: i-0c3e70929edb5d691
safekeepers:
hosts:
safekeeper-0.us-east-2.aws.neon.build:
ansible_host: i-027662bd552bf5db0
safekeeper-1.us-east-2.aws.neon.build:
ansible_host: i-0171efc3604a7b907
safekeeper-2.us-east-2.aws.neon.build:
ansible_host: i-0de0b03a51676a6ce

View File

@@ -1,5 +1,5 @@
[Unit]
Description=Zenith pageserver
Description=Neon pageserver
After=network.target auditd.service
[Service]

View File

@@ -1,12 +1,12 @@
[Unit]
Description=Zenith safekeeper
Description=Neon safekeeper
After=network.target auditd.service
[Service]
Type=simple
User=safekeeper
Environment=RUST_BACKTRACE=1 NEON_REPO_DIR=/storage/safekeeper/data LD_LIBRARY_PATH=/usr/local/v14/lib
ExecStart=/usr/local/bin/safekeeper -l {{ inventory_hostname }}.local:6500 --listen-http {{ inventory_hostname }}.local:7676 -D /storage/safekeeper/data --broker-endpoints={{ etcd_endpoints }} --remote-storage='{bucket_name="{{bucket_name}}", bucket_region="{{bucket_region}}", prefix_in_bucket="{{ env_name }}/wal"}'
ExecStart=/usr/local/bin/safekeeper -l {{ inventory_hostname }}{{ hostname_suffix }}:6500 --listen-http {{ inventory_hostname }}{{ hostname_suffix }}:7676 -D /storage/safekeeper/data --broker-endpoints={{ etcd_endpoints }} --remote-storage='{bucket_name="{{bucket_name}}", bucket_region="{{bucket_region}}", prefix_in_bucket="{{ safekeeper_s3_prefix }}"}'
ExecReload=/bin/kill -HUP $MAINPID
KillMode=mixed
KillSignal=SIGINT

View File

@@ -0,0 +1 @@
{{ pageserver_config | sivel.toiletwater.to_toml }}

View File

@@ -0,0 +1,31 @@
# Helm chart values for neon-proxy-scram.
# This is a YAML-formatted file.
image:
repository: neondatabase/neon
settings:
authBackend: "console"
authEndpoint: "http://console-staging.local/management/api/v2"
domain: "*.us-east-2.aws.neon.build"
# -- Additional labels for neon-proxy pods
podLabels:
zenith_service: proxy-scram
zenith_env: dev
zenith_region: us-east-2
zenith_region_slug: us-east-2
exposedService:
annotations:
service.beta.kubernetes.io/aws-load-balancer-type: external
service.beta.kubernetes.io/aws-load-balancer-nlb-target-type: ip
service.beta.kubernetes.io/aws-load-balancer-scheme: internet-facing
external-dns.alpha.kubernetes.io/hostname: us-east-2.aws.neon.build
#metrics:
# enabled: true
# serviceMonitor:
# enabled: true
# selector:
# release: kube-prometheus-stack

View File

@@ -46,7 +46,7 @@ jobs:
runs-on: [self-hosted, zenith-benchmarker]
env:
POSTGRES_DISTRIB_DIR: /tmp/pg_install
POSTGRES_DISTRIB_DIR: /usr/pgsql
DEFAULT_PG_VERSION: 14
steps:
@@ -138,22 +138,31 @@ jobs:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
pgbench-compare:
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
TEST_PG_BENCH_SCALES_MATRIX: "10gb"
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
DEFAULT_PG_VERSION: 14
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: remote
SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
strategy:
fail-fast: false
matrix:
# neon-captest-new: Run pgbench in a freshly created project
# neon-captest-reuse: Same, but reusing existing project
# neon-captest-prefetch: Same, with prefetching enabled (new project)
platform: [ neon-captest-new, neon-captest-reuse, neon-captest-prefetch, rds-aurora ]
platform: [ neon-captest-new, neon-captest-reuse, neon-captest-prefetch ]
db_size: [ 10gb ]
include:
- platform: neon-captest-new
db_size: 50gb
- platform: neon-captest-prefetch
db_size: 50gb
- platform: rds-aurora
db_size: 50gb
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
TEST_PG_BENCH_SCALES_MATRIX: ${{ matrix.db_size }}
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
DEFAULT_PG_VERSION: 14
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: remote
SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
PLATFORM: ${{ matrix.platform }}
runs-on: dev
container:
@@ -178,7 +187,7 @@ jobs:
echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
- name: Create Neon Project
if: matrix.platform != 'neon-captest-reuse'
if: contains(fromJson('["neon-captest-new", "neon-captest-prefetch"]'), matrix.platform)
id: create-neon-project
uses: ./.github/actions/neon-project-create
with:
@@ -204,11 +213,9 @@ jobs:
;;
esac
echo "::set-output name=connstr::${CONNSTR}"
echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
psql ${CONNSTR} -c "SELECT version();"
env:
PLATFORM: ${{ matrix.platform }}
- name: Set database options
if: matrix.platform == 'neon-captest-prefetch'
@@ -227,7 +234,6 @@ jobs:
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_init
env:
PLATFORM: ${{ matrix.platform }}
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -241,7 +247,6 @@ jobs:
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_simple_update
env:
PLATFORM: ${{ matrix.platform }}
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -255,7 +260,6 @@ jobs:
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_select_only
env:
PLATFORM: ${{ matrix.platform }}
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -268,7 +272,7 @@ jobs:
build_type: ${{ env.BUILD_TYPE }}
- name: Delete Neon Project
if: ${{ matrix.platform != 'neon-captest-reuse' && always() }}
if: ${{ steps.create-neon-project.outputs.project_id && always() }}
uses: ./.github/actions/neon-project-delete
with:
environment: dev

View File

@@ -35,12 +35,12 @@ jobs:
echo ref:$GITHUB_REF_NAME
echo rev:$(git rev-list --count HEAD)
if [[ "$GITHUB_REF_NAME" == "main" ]]; then
echo "::set-output name=tag::$(git rev-list --count HEAD)"
echo "tag=$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
elif [[ "$GITHUB_REF_NAME" == "release" ]]; then
echo "::set-output name=tag::release-$(git rev-list --count HEAD)"
echo "tag=release-$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
else
echo "GITHUB_REF_NAME (value '$GITHUB_REF_NAME') is not set to either 'main' or 'release'"
echo "::set-output name=tag::$GITHUB_RUN_ID"
echo "tag=$GITHUB_RUN_ID" >> $GITHUB_OUTPUT
fi
shell: bash
id: build-tag
@@ -78,12 +78,12 @@ jobs:
- name: Set pg 14 revision for caching
id: pg_v14_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v14)
run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v14) >> $GITHUB_OUTPUT
shell: bash -euxo pipefail {0}
- name: Set pg 15 revision for caching
id: pg_v15_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v15)
run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v15) >> $GITHUB_OUTPUT
shell: bash -euxo pipefail {0}
# Set some environment variables used by all the steps.
@@ -127,8 +127,8 @@ jobs:
target/
# Fall back to older versions of the key, if no cache for current Cargo.lock was found
key: |
v8-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
v8-${{ runner.os }}-${{ matrix.build_type }}-cargo-
v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-
- name: Cache postgres v14 build
id: cache_pg_14
@@ -389,7 +389,7 @@ jobs:
!~/.cargo/registry/src
~/.cargo/git/
target/
key: v8-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
key: v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
- name: Get Neon artifact
uses: ./.github/actions/download
@@ -481,6 +481,7 @@ jobs:
neon-image:
runs-on: dev
needs: [ tag ]
container: gcr.io/kaniko-project/executor:v1.9.0-debug
steps:
@@ -494,10 +495,11 @@ jobs:
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build neon
run: /kaniko/executor --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:$GITHUB_RUN_ID
run: /kaniko/executor --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --build-arg GIT_VERSION=${{ github.sha }} --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}}
compute-tools-image:
runs-on: dev
needs: [ tag ]
container: gcr.io/kaniko-project/executor:v1.9.0-debug
steps:
@@ -508,11 +510,12 @@ jobs:
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute tools
run: /kaniko/executor --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --dockerfile Dockerfile.compute-tools --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:$GITHUB_RUN_ID
run: /kaniko/executor --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --build-arg GIT_VERSION=${{ github.sha }} --dockerfile Dockerfile.compute-tools --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{needs.tag.outputs.build-tag}}
compute-node-image:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
needs: [ tag ]
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
@@ -527,11 +530,12 @@ jobs:
# cloud repo depends on this image name, thus duplicating it
# remove compute-node when cloud repo is updated
- name: Kaniko build compute node with extensions v14 (compatibility)
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:$GITHUB_RUN_ID
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --build-arg GIT_VERSION=${{ github.sha }} --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:${{needs.tag.outputs.build-tag}}
compute-node-image-v14:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
needs: [ tag ]
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
@@ -543,12 +547,13 @@ jobs:
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute node with extensions v14
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:$GITHUB_RUN_ID
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --build-arg GIT_VERSION=${{ github.sha }} --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:${{needs.tag.outputs.build-tag}}
compute-node-image-v15:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
needs: [ tag ]
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
@@ -560,11 +565,11 @@ jobs:
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute node with extensions v15
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --dockerfile Dockerfile.compute-node-v15 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:$GITHUB_RUN_ID
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --build-arg GIT_VERSION=${{ github.sha }} --dockerfile Dockerfile.compute-node-v15 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:${{needs.tag.outputs.build-tag}}
promote-images:
runs-on: dev
needs: [ neon-image, compute-node-image, compute-node-image-v14, compute-tools-image ]
needs: [ tag, neon-image, compute-node-image, compute-node-image-v14, compute-node-image-v15, compute-tools-image ]
if: github.event_name != 'workflow_dispatch'
container: amazon/aws-cli
strategy:
@@ -573,12 +578,13 @@ jobs:
# compute-node uses postgres 14, which is default now
# cloud repo depends on this image name, thus duplicating it
# remove compute-node when cloud repo is updated
name: [ neon, compute-node, compute-node-v14, compute-tools ]
name: [ neon, compute-node, compute-node-v14, compute-node-v15, compute-tools ]
steps:
- name: Promote image to latest
run:
MANIFEST=$(aws ecr batch-get-image --repository-name ${{ matrix.name }} --image-ids imageTag=$GITHUB_RUN_ID --query 'images[].imageManifest' --output text) && aws ecr put-image --repository-name ${{ matrix.name }} --image-tag latest --image-manifest "$MANIFEST"
run: |
export MANIFEST=$(aws ecr batch-get-image --repository-name ${{ matrix.name }} --image-ids imageTag=${{needs.tag.outputs.build-tag}} --query 'images[].imageManifest' --output text)
aws ecr put-image --repository-name ${{ matrix.name }} --image-tag latest --image-manifest "$MANIFEST"
push-docker-hub:
runs-on: dev
@@ -597,16 +603,19 @@ jobs:
echo "{\"credsStore\":\"ecr-login\"}" > /github/home/.docker/config.json
- name: Pull neon image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:latest neon
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}} neon
- name: Pull compute tools image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:latest compute-tools
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{needs.tag.outputs.build-tag}} compute-tools
- name: Pull compute node image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:latest compute-node
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:${{needs.tag.outputs.build-tag}} compute-node
- name: Pull compute node v14 image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:latest compute-node-v14
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:${{needs.tag.outputs.build-tag}} compute-node-v14
- name: Pull compute node v15 image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:${{needs.tag.outputs.build-tag}} compute-node-v15
- name: Pull rust image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned rust
@@ -616,9 +625,11 @@ jobs:
(github.ref_name == 'main' || github.ref_name == 'release') &&
github.event_name != 'workflow_dispatch'
run: |
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/neon:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-tools:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-node:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.us-east-2.amazonaws.com/neon:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-tools:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-node:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-node-v14:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-node-v15:latest
- name: Configure Docker Hub login
run: |
@@ -638,6 +649,9 @@ jobs:
- name: Push compute node v14 image to Docker Hub
run: crane push compute-node-v14 neondatabase/compute-node-v14:${{needs.tag.outputs.build-tag}}
- name: Push compute node v15 image to Docker Hub
run: crane push compute-node-v15 neondatabase/compute-node-v15:${{needs.tag.outputs.build-tag}}
- name: Push rust image to Docker Hub
run: crane push rust neondatabase/rust:pinned
@@ -650,6 +664,7 @@ jobs:
crane tag neondatabase/compute-tools:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v15:${{needs.tag.outputs.build-tag}} latest
calculate-deploy-targets:
runs-on: [ self-hosted, Linux, k8s-runner ]
@@ -662,12 +677,12 @@ jobs:
- id: set-matrix
run: |
if [[ "$GITHUB_REF_NAME" == "main" ]]; then
STAGING='{"env_name": "staging", "proxy_job": "neon-proxy", "proxy_config": "staging.proxy", "kubeconfig_secret": "STAGING_KUBECONFIG_DATA"}'
NEON_STRESS='{"env_name": "neon-stress", "proxy_job": "neon-stress-proxy", "proxy_config": "neon-stress.proxy", "kubeconfig_secret": "NEON_STRESS_KUBECONFIG_DATA"}'
echo "::set-output name=include::[$STAGING, $NEON_STRESS]"
STAGING='{"env_name": "staging", "proxy_job": "neon-proxy", "proxy_config": "staging.proxy", "kubeconfig_secret": "STAGING_KUBECONFIG_DATA", "console_api_key_secret": "NEON_STAGING_API_KEY"}'
NEON_STRESS='{"env_name": "neon-stress", "proxy_job": "neon-stress-proxy", "proxy_config": "neon-stress.proxy", "kubeconfig_secret": "NEON_STRESS_KUBECONFIG_DATA", "console_api_key_secret": "NEON_CAPTEST_API_KEY"}'
echo "include=[$STAGING, $NEON_STRESS]" >> $GITHUB_OUTPUT
elif [[ "$GITHUB_REF_NAME" == "release" ]]; then
PRODUCTION='{"env_name": "production", "proxy_job": "neon-proxy", "proxy_config": "production.proxy", "kubeconfig_secret": "PRODUCTION_KUBECONFIG_DATA"}'
echo "::set-output name=include::[$PRODUCTION]"
PRODUCTION='{"env_name": "production", "proxy_job": "neon-proxy", "proxy_config": "production.proxy", "kubeconfig_secret": "PRODUCTION_KUBECONFIG_DATA", "console_api_key_secret": "NEON_PRODUCTION_API_KEY"}'
echo "include=[$PRODUCTION]" >> $GITHUB_OUTPUT
else
echo "GITHUB_REF_NAME (value '$GITHUB_REF_NAME') is not set to either 'main' or 'release'"
exit 1
@@ -703,7 +718,7 @@ jobs:
- name: Setup ansible
run: |
export PATH="/root/.local/bin:$PATH"
pip install --progress-bar off --user ansible boto3
pip install --progress-bar off --user ansible boto3 toml
- name: Redeploy
run: |
@@ -725,8 +740,48 @@ jobs:
chmod 0600 ssh-key
ssh-add ssh-key
rm -f ssh-key ssh-key-cert.pub
ansible-galaxy collection install sivel.toiletwater
ansible-playbook deploy.yaml -i ${{ matrix.env_name }}.hosts.yaml -e CONSOLE_API_TOKEN=${{ secrets[matrix.console_api_key_secret] }}
rm -f neon_install.tar.gz .neon_current_version
ansible-playbook deploy.yaml -i ${{ matrix.env_name }}.hosts
deploy-new:
runs-on: dev
container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/ansible:pinned
# We need both storage **and** compute images for deploy, because control plane picks the compute version based on the storage version.
# If it notices a fresh storage it may bump the compute version. And if compute image failed to build it may break things badly
needs: [ push-docker-hub, calculate-deploy-targets, tag, regress-tests ]
if: |
(github.ref_name == 'main') &&
github.event_name != 'workflow_dispatch'
defaults:
run:
shell: bash
env:
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
steps:
- name: Checkout
uses: actions/checkout@v3
with:
submodules: true
fetch-depth: 0
- name: Redeploy
run: |
export DOCKER_TAG=${{needs.tag.outputs.build-tag}}
cd "$(pwd)/.github/ansible"
if [[ "$GITHUB_REF_NAME" == "main" ]]; then
./get_binaries.sh
elif [[ "$GITHUB_REF_NAME" == "release" ]]; then
RELEASE=true ./get_binaries.sh
else
echo "GITHUB_REF_NAME (value '$GITHUB_REF_NAME') is not set to either 'main' or 'release'"
exit 1
fi
ansible-galaxy collection install sivel.toiletwater
ansible-playbook deploy.yaml -i staging.us-east-2.hosts.yaml -e @ssm_config -e CONSOLE_API_TOKEN=${{secrets.NEON_STAGING_API_KEY}}
rm -f neon_install.tar.gz .neon_current_version
deploy-proxy:
@@ -768,5 +823,54 @@ jobs:
- name: Re-deploy proxy
run: |
DOCKER_TAG=${{needs.tag.outputs.build-tag}}
helm upgrade ${{ matrix.proxy_job }} neondatabase/neon-proxy --namespace default --install -f .github/helm-values/${{ matrix.proxy_config }}.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }}-scram neondatabase/neon-proxy --namespace default --install -f .github/helm-values/${{ matrix.proxy_config }}-scram.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }} neondatabase/neon-proxy --namespace neon-proxy --install -f .github/helm-values/${{ matrix.proxy_config }}.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }}-scram neondatabase/neon-proxy --namespace neon-proxy --install -f .github/helm-values/${{ matrix.proxy_config }}-scram.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
deploy-proxy-new:
runs-on: dev
container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/ansible:pinned
# Compute image isn't strictly required for proxy deploy, but let's still wait for it to run all deploy jobs consistently.
needs: [ push-docker-hub, calculate-deploy-targets, tag, regress-tests ]
if: |
(github.ref_name == 'main') &&
github.event_name != 'workflow_dispatch'
defaults:
run:
shell: bash
steps:
- name: Checkout
uses: actions/checkout@v3
with:
submodules: true
fetch-depth: 0
- name: Configure environment
run: |
helm repo add neondatabase https://neondatabase.github.io/helm-charts
aws --region us-east-2 eks update-kubeconfig --name dev-us-east-2-beta --role-arn arn:aws:iam::369495373322:role/github-runner
- name: Re-deploy proxy
run: |
DOCKER_TAG=${{needs.tag.outputs.build-tag}}
helm upgrade neon-proxy-scram neondatabase/neon-proxy --namespace neon-proxy --create-namespace --install -f .github/helm-values/dev-us-east-2-beta.neon-proxy-scram.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
promote-compatibility-test-snapshot:
runs-on: dev
container:
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
options: --init
needs: [ deploy, deploy-proxy ]
if: github.ref_name == 'release' && github.event_name != 'workflow_dispatch'
steps:
- name: Promote compatibility snapshot for the release
shell: bash -euxo pipefail {0}
env:
BUCKET: neon-github-public-dev
PREFIX: artifacts/latest
run: |
for build_type in debug release; do
OLD_FILENAME=compatibility-snapshot-${build_type}-pg14-${GITHUB_RUN_ID}.tar.zst
NEW_FILENAME=compatibility-snapshot-${build_type}-pg14.tar.zst
time aws s3 mv --only-show-errors s3://${BUCKET}/${PREFIX}/${OLD_FILENAME} s3://${BUCKET}/${PREFIX}/${NEW_FILENAME}
done

View File

@@ -36,7 +36,7 @@ jobs:
steps:
- name: Checkout
uses: actions/checkout@v2
uses: actions/checkout@v3
with:
submodules: true
fetch-depth: 2
@@ -56,12 +56,12 @@ jobs:
- name: Set pg 14 revision for caching
id: pg_v14_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v14)
run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v14) >> $GITHUB_OUTPUT
shell: bash -euxo pipefail {0}
- name: Set pg 15 revision for caching
id: pg_v15_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v15)
run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v15) >> $GITHUB_OUTPUT
shell: bash -euxo pipefail {0}
- name: Cache postgres v14 build
@@ -106,7 +106,7 @@ jobs:
!~/.cargo/registry/src
~/.cargo/git
target
key: v4-${{ runner.os }}-cargo-${{ hashFiles('./Cargo.lock') }}-rust
key: v5-${{ runner.os }}-cargo-${{ hashFiles('./Cargo.lock') }}-rust
- name: Run cargo clippy
run: ./run_clippy.sh

987
Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -1,3 +1,14 @@
# 'named-profiles' feature was stabilized in cargo 1.57. This line makes the
# build work with older cargo versions.
#
# We have this because as of this writing, the latest cargo Debian package
# that's available is 1.56. (Confusingly, the Debian package version number
# is 0.57, whereas 'cargo --version' says 1.56.)
#
# See https://tracker.debian.org/pkg/cargo for the current status of the
# package. When that gets updated, we can remove this.
cargo-features = ["named-profiles"]
[workspace]
members = [
"compute_tools",

View File

@@ -44,7 +44,7 @@ COPY . .
# Show build caching stats to check if it was used in the end.
# Has to be the part of the same RUN since cachepot daemon is killed in the end of this RUN, losing the compilation stats.
RUN set -e \
&& mold -run cargo build --bin pageserver --bin safekeeper --bin proxy --locked --release \
&& mold -run cargo build --bin pageserver --bin pageserver_binutils --bin draw_timeline_dir --bin safekeeper --bin proxy --locked --release \
&& cachepot -s
# Build final image
@@ -63,9 +63,11 @@ RUN set -e \
&& useradd -d /data neon \
&& chown -R neon:neon /data
COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/safekeeper /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/proxy /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver_binutils /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/draw_timeline_dir /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/safekeeper /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/proxy /usr/local/bin
COPY --from=pg-build /home/nonroot/pg_install/v14 /usr/local/v14/
COPY --from=pg-build /home/nonroot/pg_install/v15 /usr/local/v15/
@@ -85,4 +87,3 @@ VOLUME ["/data"]
USER neon
EXPOSE 6400
EXPOSE 9898
CMD ["/bin/bash"]

View File

@@ -1,24 +1,26 @@
ARG TAG=pinned
# apparently, ARGs don't get replaced in RUN commands in kaniko
# ARG POSTGIS_VERSION=3.3.0
# ARG PLV8_VERSION=3.1.4
# ARG PG_VERSION=v14
#
# This file is identical to the Dockerfile.compute-node-v15 file
# except for the version of Postgres that is built.
#
ARG TAG=pinned
#########################################################################################
#
# Layer "build-deps"
#
#########################################################################################
FROM debian:bullseye-slim AS build-deps
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update
RUN apt update && \
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev wget pkg-config libglib2.0-dev
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev \
zlib1g-dev libxml2-dev libcurl4-openssl-dev libossp-uuid-dev wget pkg-config
#########################################################################################
#
# Layer "pg-build"
# Build Postgres from the neon postgres repository.
#
#########################################################################################
FROM build-deps AS pg-build
COPY vendor/postgres-v14 postgres
RUN cd postgres && \
@@ -29,22 +31,20 @@ RUN cd postgres && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/include install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/interfaces/libpq install
#########################################################################################
#
# Layer "postgis-build"
# Build PostGIS from the upstream PostGIS mirror.
#
# PostGIS compiles against neon postgres sources without changes. Perhaps we
# could even use the upstream binaries, compiled against vanilla Postgres, but
# it would require some investigation to check that it works, and also keeps
# working in the future. So for now, we compile our own binaries.
#########################################################################################
FROM build-deps AS postgis-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
tar xvzf postgis-3.3.0.tar.gz && \
cd postgis-3.3.0 && \
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.1.tar.gz && \
tar xvzf postgis-3.3.1.tar.gz && \
cd postgis-3.3.1 && \
./autogen.sh && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
./configure && \
@@ -57,39 +57,55 @@ RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_tiger_geocoder.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_topology.control
#########################################################################################
#
# Layer "plv8-build"
# Build plv8
#
#########################################################################################
FROM build-deps AS plv8-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y ninja-build python3-dev libc++-dev libc++abi-dev libncurses5
apt install -y ninja-build python3-dev libc++-dev libc++abi-dev libncurses5 binutils
# https://github.com/plv8/plv8/issues/475
# Debian bullseye provides binutils 2.35 when >= 2.38 is necessary
RUN apt update && \
apt install -y --no-install-recommends -t testing binutils
# https://github.com/plv8/plv8/issues/475:
# v8 uses gold for linking and sets `--thread-count=4` which breaks
# gold version <= 1.35 (https://sourceware.org/bugzilla/show_bug.cgi?id=23607)
# Install newer gold version manually as debian-testing binutils version updates
# libc version, which in turn breaks other extension built against non-testing libc.
RUN wget https://ftp.gnu.org/gnu/binutils/binutils-2.38.tar.gz && \
tar xvzf binutils-2.38.tar.gz && \
cd binutils-2.38 && \
cd libiberty && ./configure && make -j $(getconf _NPROCESSORS_ONLN) && \
cd ../bfd && ./configure && make bfdver.h && \
cd ../gold && ./configure && make -j $(getconf _NPROCESSORS_ONLN) && make install && \
cp /usr/local/bin/ld.gold /usr/bin/gold
# Sed is used to patch for https://github.com/plv8/plv8/issues/503
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.4.tar.gz && \
tar xvzf v3.1.4.tar.gz && \
cd plv8-3.1.4 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
sed -i 's/MemoryContextAlloc(/MemoryContextAllocZero(/' plv8.cc && \
make DOCKER=1 -j $(getconf _NPROCESSORS_ONLN) install && \
rm -rf /plv8-* && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plv8.control
#########################################################################################
#
# Layer "h3-pg-build"
# Build h3_pg
#
#########################################################################################
FROM build-deps AS h3-pg-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
# packaged cmake is too old
RUN apt update && \
apt install -y --no-install-recommends -t testing cmake
RUN wget https://github.com/Kitware/CMake/releases/download/v3.24.2/cmake-3.24.2-linux-x86_64.sh \
-q -O /tmp/cmake-install.sh \
&& chmod u+x /tmp/cmake-install.sh \
&& /tmp/cmake-install.sh --skip-license --prefix=/usr/local/ \
&& rm /tmp/cmake-install.sh
RUN wget https://github.com/uber/h3/archive/refs/tags/v4.0.1.tar.gz -O h3.tgz && \
tar xvzf h3.tgz && \
@@ -108,16 +124,18 @@ RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.0.1.tar.gz -O h3
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3.control
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3_postgis.control
#########################################################################################
#
# Layer "neon-pg-ext-build"
# compile neon extensions
#
#########################################################################################
FROM build-deps AS neon-pg-ext-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
# plv8 still sometimes crashes during the creation
# COPY --from=plv8-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=plv8-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /h3/usr /
COPY pgxn/ pgxn/
@@ -127,16 +145,22 @@ RUN make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon \
-s install
#########################################################################################
#
# Compile and run the Neon-specific `compute_ctl` binary
#
#########################################################################################
FROM 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:$TAG AS compute-tools
USER nonroot
# Copy entire project to get Cargo.* files with proper dependencies for the whole project
COPY --chown=nonroot . .
RUN cd compute_tools && cargo build --locked --profile release-line-debug-size-lto
#########################################################################################
#
# Clean up postgres folder before inclusion
#
#########################################################################################
FROM neon-pg-ext-build AS postgres-cleanup-layer
COPY --from=neon-pg-ext-build /usr/local/pgsql /usr/local/pgsql
@@ -154,10 +178,12 @@ RUN rm -r /usr/local/pgsql/lib/pgxs/src
# if they were to be used by other libraries.
RUN rm /usr/local/pgsql/lib/lib*.a
#########################################################################################
#
# Final layer
# Put it all together into the final image
#
#########################################################################################
FROM debian:bullseye-slim
# Add user postgres
RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
@@ -174,8 +200,6 @@ COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-deb
# libreadline8 for psql
# libossp-uuid16 for extension ossp-uuid
# libgeos, libgdal, libproj and libprotobuf-c1 for PostGIS
# GLIBC 2.34 for plv8.
# Debian bullseye provides GLIBC 2.31, so we install the library from testing
#
# Lastly, link compute_ctl into zenith_ctl while we're at it,
# so that we don't need to put this in another layer.
@@ -188,12 +212,6 @@ RUN apt update && \
libproj19 \
libprotobuf-c1 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
echo "Installing GLIBC 2.34" && \
echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing libc6 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
ln /usr/local/bin/compute_ctl /usr/local/bin/zenith_ctl
USER postgres

View File

@@ -4,23 +4,23 @@
#
ARG TAG=pinned
# apparently, ARGs don't get replaced in RUN commands in kaniko
# ARG POSTGIS_VERSION=3.3.0
# ARG PLV8_VERSION=3.1.4
# ARG PG_VERSION=v15
#########################################################################################
#
# Layer "build-deps"
#
#########################################################################################
FROM debian:bullseye-slim AS build-deps
RUN apt update && \
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev \
zlib1g-dev libxml2-dev libcurl4-openssl-dev libossp-uuid-dev wget pkg-config
#########################################################################################
#
# Layer "pg-build"
# Build Postgres from the neon postgres repository.
#
#########################################################################################
FROM build-deps AS pg-build
COPY vendor/postgres-v15 postgres
RUN cd postgres && \
@@ -31,22 +31,20 @@ RUN cd postgres && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/include install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/interfaces/libpq install
#########################################################################################
#
# Layer "postgis-build"
# Build PostGIS from the upstream PostGIS mirror.
#
# PostGIS compiles against neon postgres sources without changes. Perhaps we
# could even use the upstream binaries, compiled against vanilla Postgres, but
# it would require some investigation to check that it works, and also keeps
# working in the future. So for now, we compile our own binaries.
#########################################################################################
FROM build-deps AS postgis-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc wget
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
tar xvzf postgis-3.3.0.tar.gz && \
cd postgis-3.3.0 && \
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.1.tar.gz && \
tar xvzf postgis-3.3.1.tar.gz && \
cd postgis-3.3.1 && \
./autogen.sh && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
./configure && \
@@ -59,37 +57,87 @@ RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_tiger_geocoder.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_topology.control
#########################################################################################
#
# Layer "plv8-build"
# Build plv8
#
#########################################################################################
FROM build-deps AS plv8-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y git curl wget make ninja-build build-essential libncurses5 python3-dev pkg-config libc++-dev libc++abi-dev libglib2.0-dev
apt install -y ninja-build python3-dev libc++-dev libc++abi-dev libncurses5 binutils
# https://github.com/plv8/plv8/issues/475
# Debian bullseye provides binutils 2.35 when >= 2.38 is necessary
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing binutils
# https://github.com/plv8/plv8/issues/475:
# v8 uses gold for linking and sets `--thread-count=4` which breaks
# gold version <= 1.35 (https://sourceware.org/bugzilla/show_bug.cgi?id=23607)
# Install newer gold version manually as debian-testing binutils version updates
# libc version, which in turn breaks other extension built against non-testing libc.
RUN wget https://ftp.gnu.org/gnu/binutils/binutils-2.38.tar.gz && \
tar xvzf binutils-2.38.tar.gz && \
cd binutils-2.38 && \
cd libiberty && ./configure && make -j $(getconf _NPROCESSORS_ONLN) && \
cd ../bfd && ./configure && make bfdver.h && \
cd ../gold && ./configure && make -j $(getconf _NPROCESSORS_ONLN) && make install && \
cp /usr/local/bin/ld.gold /usr/bin/gold
# Sed is used to patch for https://github.com/plv8/plv8/issues/503
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.4.tar.gz && \
tar xvzf v3.1.4.tar.gz && \
cd plv8-3.1.4 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
sed -i 's/MemoryContextAlloc(/MemoryContextAllocZero(/' plv8.cc && \
make DOCKER=1 -j $(getconf _NPROCESSORS_ONLN) install && \
rm -rf /plv8-* && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plv8.control
#########################################################################################
#
# Layer "h3-pg-build"
# Build h3_pg
#
#########################################################################################
FROM build-deps AS h3-pg-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
# packaged cmake is too old
RUN wget https://github.com/Kitware/CMake/releases/download/v3.24.2/cmake-3.24.2-linux-x86_64.sh \
-q -O /tmp/cmake-install.sh \
&& chmod u+x /tmp/cmake-install.sh \
&& /tmp/cmake-install.sh --skip-license --prefix=/usr/local/ \
&& rm /tmp/cmake-install.sh
RUN wget https://github.com/uber/h3/archive/refs/tags/v4.0.1.tar.gz -O h3.tgz && \
tar xvzf h3.tgz && \
cd h3-4.0.1 && \
mkdir build && \
cd build && \
cmake .. -DCMAKE_BUILD_TYPE=Release && \
make -j $(getconf _NPROCESSORS_ONLN) && \
DESTDIR=/h3 make install && \
cp -R /h3/usr / && \
rm -rf build
RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.0.1.tar.gz -O h3-pg.tgz && \
tar xvzf h3-pg.tgz && \
cd h3-pg-4.0.1 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3_postgis.control
#########################################################################################
#
# Layer "neon-pg-ext-build"
# compile neon extensions
#
#########################################################################################
FROM build-deps AS neon-pg-ext-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=plv8-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /h3/usr /
COPY pgxn/ pgxn/
RUN make -j $(getconf _NPROCESSORS_ONLN) \
@@ -97,16 +145,22 @@ RUN make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon \
-s install
#########################################################################################
#
# Compile and run the Neon-specific `compute_ctl` binary
#
#########################################################################################
FROM 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:$TAG AS compute-tools
USER nonroot
# Copy entire project to get Cargo.* files with proper dependencies for the whole project
COPY --chown=nonroot . .
RUN cd compute_tools && cargo build --locked --profile release-line-debug-size-lto
#########################################################################################
#
# Clean up postgres folder before inclusion
#
#########################################################################################
FROM neon-pg-ext-build AS postgres-cleanup-layer
COPY --from=neon-pg-ext-build /usr/local/pgsql /usr/local/pgsql
@@ -124,10 +178,12 @@ RUN rm -r /usr/local/pgsql/lib/pgxs/src
# if they were to be used by other libraries.
RUN rm /usr/local/pgsql/lib/lib*.a
#########################################################################################
#
# Final layer
# Put it all together into the final image
#
#########################################################################################
FROM debian:bullseye-slim
# Add user postgres
RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
@@ -137,8 +193,6 @@ RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
chmod 0750 /var/db/postgres/compute && \
echo '/usr/local/lib' >> /etc/ld.so.conf && /sbin/ldconfig
# TODO: Check if we can make the extension setup more modular versus a linear build
# currently plv8-build copies the output /usr/local/pgsql from postgis-build, etc#
COPY --from=postgres-cleanup-layer --chown=postgres /usr/local/pgsql /usr/local
COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-debug-size-lto/compute_ctl /usr/local/bin/compute_ctl
@@ -146,8 +200,6 @@ COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-deb
# libreadline8 for psql
# libossp-uuid16 for extension ossp-uuid
# libgeos, libgdal, libproj and libprotobuf-c1 for PostGIS
# GLIBC 2.34 for plv8.
# Debian bullseye provides GLIBC 2.31, so we install the library from testing
#
# Lastly, link compute_ctl into zenith_ctl while we're at it,
# so that we don't need to put this in another layer.
@@ -160,12 +212,6 @@ RUN apt update && \
libproj19 \
libprotobuf-c1 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
echo "Installing GLIBC 2.34" && \
echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing libc6 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
ln /usr/local/bin/compute_ctl /usr/local/bin/zenith_ctl
USER postgres

View File

@@ -6,10 +6,12 @@ edition = "2021"
[dependencies]
anyhow = "1.0"
chrono = "0.4"
clap = "3.0"
clap = "4.0"
env_logger = "0.9"
futures = "0.3.13"
hyper = { version = "0.14", features = ["full"] }
log = { version = "0.4", features = ["std", "serde"] }
notify = "5.0.0"
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
regex = "1"
serde = { version = "1.0", features = ["derive"] }

View File

@@ -51,53 +51,19 @@ fn main() -> Result<()> {
// TODO: re-use `utils::logging` later
init_logger(DEFAULT_LOG_LEVEL)?;
// Env variable is set by `cargo`
let version: Option<&str> = option_env!("CARGO_PKG_VERSION");
let matches = clap::App::new("compute_ctl")
.version(version.unwrap_or("unknown"))
.arg(
Arg::new("connstr")
.short('C')
.long("connstr")
.value_name("DATABASE_URL")
.required(true),
)
.arg(
Arg::new("pgdata")
.short('D')
.long("pgdata")
.value_name("DATADIR")
.required(true),
)
.arg(
Arg::new("pgbin")
.short('b')
.long("pgbin")
.value_name("POSTGRES_PATH"),
)
.arg(
Arg::new("spec")
.short('s')
.long("spec")
.value_name("SPEC_JSON"),
)
.arg(
Arg::new("spec-path")
.short('S')
.long("spec-path")
.value_name("SPEC_PATH"),
)
.get_matches();
let matches = cli().get_matches();
let pgdata = matches.value_of("pgdata").expect("PGDATA path is required");
let pgdata = matches
.get_one::<String>("pgdata")
.expect("PGDATA path is required");
let connstr = matches
.value_of("connstr")
.get_one::<String>("connstr")
.expect("Postgres connection string is required");
let spec = matches.value_of("spec");
let spec_path = matches.value_of("spec-path");
let spec = matches.get_one::<String>("spec");
let spec_path = matches.get_one::<String>("spec-path");
// Try to use just 'postgres' if no path is provided
let pgbin = matches.value_of("pgbin").unwrap_or("postgres");
let pgbin = matches.get_one::<String>("pgbin").unwrap();
let spec: ComputeSpec = match spec {
// First, try to get cluster spec from the cli argument
@@ -173,3 +139,48 @@ fn main() -> Result<()> {
}
}
}
fn cli() -> clap::Command {
// Env variable is set by `cargo`
let version = option_env!("CARGO_PKG_VERSION").unwrap_or("unknown");
clap::Command::new("compute_ctl")
.version(version)
.arg(
Arg::new("connstr")
.short('C')
.long("connstr")
.value_name("DATABASE_URL")
.required(true),
)
.arg(
Arg::new("pgdata")
.short('D')
.long("pgdata")
.value_name("DATADIR")
.required(true),
)
.arg(
Arg::new("pgbin")
.short('b')
.long("pgbin")
.default_value("postgres")
.value_name("POSTGRES_PATH"),
)
.arg(
Arg::new("spec")
.short('s')
.long("spec")
.value_name("SPEC_JSON"),
)
.arg(
Arg::new("spec-path")
.short('S')
.long("spec-path")
.value_name("SPEC_PATH"),
)
}
#[test]
fn verify_cli() {
cli().debug_assert()
}

View File

@@ -178,7 +178,6 @@ impl ComputeNode {
.args(&["--sync-safekeepers"])
.env("PGDATA", &self.pgdata) // we cannot use -D in this mode
.stdout(Stdio::piped())
.stderr(Stdio::piped())
.spawn()
.expect("postgres --sync-safekeepers failed to start");
@@ -191,10 +190,10 @@ impl ComputeNode {
if !sync_output.status.success() {
anyhow::bail!(
"postgres --sync-safekeepers exited with non-zero status: {}. stdout: {}, stderr: {}",
"postgres --sync-safekeepers exited with non-zero status: {}. stdout: {}",
sync_output.status,
String::from_utf8(sync_output.stdout).expect("postgres --sync-safekeepers exited, and stdout is not utf-8"),
String::from_utf8(sync_output.stderr).expect("postgres --sync-safekeepers exited, and stderr is not utf-8"),
String::from_utf8(sync_output.stdout)
.expect("postgres --sync-safekeepers exited, and stdout is not utf-8"),
);
}
@@ -258,14 +257,7 @@ impl ComputeNode {
.spawn()
.expect("cannot start postgres process");
// Try default Postgres port if it is not provided
let port = self
.spec
.cluster
.settings
.find("port")
.unwrap_or_else(|| "5432".to_string());
wait_for_postgres(&mut pg, &port, pgdata_path)?;
wait_for_postgres(&mut pg, pgdata_path)?;
// If connection fails,
// it may be the old node with `zenith_admin` superuser.

View File

@@ -1,18 +1,18 @@
use std::fmt::Write;
use std::fs;
use std::fs::File;
use std::io::{BufRead, BufReader};
use std::net::{SocketAddr, TcpStream};
use std::os::unix::fs::PermissionsExt;
use std::path::Path;
use std::process::Child;
use std::str::FromStr;
use std::{fs, thread, time};
use std::time::{Duration, Instant};
use anyhow::{bail, Result};
use notify::{RecursiveMode, Watcher};
use postgres::{Client, Transaction};
use serde::Deserialize;
const POSTGRES_WAIT_TIMEOUT: u64 = 60 * 1000; // milliseconds
const POSTGRES_WAIT_TIMEOUT: Duration = Duration::from_millis(60 * 1000); // milliseconds
/// Rust representation of Postgres role info with only those fields
/// that matter for us.
@@ -168,7 +168,7 @@ impl Database {
/// it may require a proper quoting too.
pub fn to_pg_options(&self) -> String {
let mut params: String = self.options.as_pg_options();
write!(params, " OWNER {}", &self.owner.quote())
write!(params, " OWNER {}", &self.owner.pg_quote())
.expect("String is documented to not to error during write operations");
params
@@ -179,18 +179,17 @@ impl Database {
/// intended to be used for DB / role names.
pub type PgIdent = String;
/// Generic trait used to provide quoting for strings used in the
/// Postgres SQL queries. Currently used only to implement quoting
/// of identifiers, but could be used for literals in the future.
pub trait PgQuote {
fn quote(&self) -> String;
/// Generic trait used to provide quoting / encoding for strings used in the
/// Postgres SQL queries and DATABASE_URL.
pub trait Escaping {
fn pg_quote(&self) -> String;
}
impl PgQuote for PgIdent {
impl Escaping for PgIdent {
/// This is intended to mimic Postgres quote_ident(), but for simplicity it
/// always quotes provided string with `""` and escapes every `"`. Not idempotent,
/// i.e. if string is already escaped it will be escaped again.
fn quote(&self) -> String {
/// always quotes provided string with `""` and escapes every `"`.
/// **Not idempotent**, i.e. if string is already escaped it will be escaped again.
fn pg_quote(&self) -> String {
let result = format!("\"{}\"", self.replace('"', "\"\""));
result
}
@@ -230,52 +229,112 @@ pub fn get_existing_dbs(client: &mut Client) -> Result<Vec<Database>> {
Ok(postgres_dbs)
}
/// Wait for Postgres to become ready to accept connections:
/// - state should be `ready` in the `pgdata/postmaster.pid`
/// - and we should be able to connect to 127.0.0.1:5432
pub fn wait_for_postgres(pg: &mut Child, port: &str, pgdata: &Path) -> Result<()> {
/// Wait for Postgres to become ready to accept connections. It's ready to
/// accept connections when the state-field in `pgdata/postmaster.pid` says
/// 'ready'.
pub fn wait_for_postgres(pg: &mut Child, pgdata: &Path) -> Result<()> {
let pid_path = pgdata.join("postmaster.pid");
let mut slept: u64 = 0; // ms
let pause = time::Duration::from_millis(100);
let timeout = time::Duration::from_millis(10);
let addr = SocketAddr::from_str(&format!("127.0.0.1:{}", port)).unwrap();
// PostgreSQL writes line "ready" to the postmaster.pid file, when it has
// completed initialization and is ready to accept connections. We want to
// react quickly and perform the rest of our initialization as soon as
// PostgreSQL starts accepting connections. Use 'notify' to be notified
// whenever the PID file is changed, and whenever it changes, read it to
// check if it's now "ready".
//
// You cannot actually watch a file before it exists, so we first watch the
// data directory, and once the postmaster.pid file appears, we switch to
// watch the file instead. We also wake up every 100 ms to poll, just in
// case we miss some events for some reason. Not strictly necessary, but
// better safe than sorry.
let (tx, rx) = std::sync::mpsc::channel();
let (mut watcher, rx): (Box<dyn Watcher>, _) = match notify::recommended_watcher(move |res| {
let _ = tx.send(res);
}) {
Ok(watcher) => (Box::new(watcher), rx),
Err(e) => {
match e.kind {
notify::ErrorKind::Io(os) if os.raw_os_error() == Some(38) => {
// docker on m1 macs does not support recommended_watcher
// but return "Function not implemented (os error 38)"
// see https://github.com/notify-rs/notify/issues/423
let (tx, rx) = std::sync::mpsc::channel();
loop {
// Sleep POSTGRES_WAIT_TIMEOUT at max (a bit longer actually if consider a TCP timeout,
// but postgres starts listening almost immediately, even if it is not really
// ready to accept connections).
if slept >= POSTGRES_WAIT_TIMEOUT {
bail!("timed out while waiting for Postgres to start");
// let's poll it faster than what we check the results for (100ms)
let config =
notify::Config::default().with_poll_interval(Duration::from_millis(50));
let watcher = notify::PollWatcher::new(
move |res| {
let _ = tx.send(res);
},
config,
)?;
(Box::new(watcher), rx)
}
_ => return Err(e.into()),
}
}
};
watcher.watch(pgdata, RecursiveMode::NonRecursive)?;
let started_at = Instant::now();
let mut postmaster_pid_seen = false;
loop {
if let Ok(Some(status)) = pg.try_wait() {
// Postgres exited, that is not what we expected, bail out earlier.
let code = status.code().unwrap_or(-1);
bail!("Postgres exited unexpectedly with code {}", code);
}
let res = rx.recv_timeout(Duration::from_millis(100));
log::debug!("woken up by notify: {res:?}");
// If there are multiple events in the channel already, we only need to be
// check once. Swallow the extra events before we go ahead to check the
// pid file.
while let Ok(res) = rx.try_recv() {
log::debug!("swallowing extra event: {res:?}");
}
// Check that we can open pid file first.
if let Ok(file) = File::open(&pid_path) {
if !postmaster_pid_seen {
log::debug!("postmaster.pid appeared");
watcher
.unwatch(pgdata)
.expect("Failed to remove pgdata dir watch");
watcher
.watch(&pid_path, RecursiveMode::NonRecursive)
.expect("Failed to add postmaster.pid file watch");
postmaster_pid_seen = true;
}
let file = BufReader::new(file);
let last_line = file.lines().last();
// Pid file could be there and we could read it, but it could be empty, for example.
if let Some(Ok(line)) = last_line {
let status = line.trim();
let can_connect = TcpStream::connect_timeout(&addr, timeout).is_ok();
log::debug!("last line of postmaster.pid: {status:?}");
// Now Postgres is ready to accept connections
if status == "ready" && can_connect {
if status == "ready" {
break;
}
}
}
thread::sleep(pause);
slept += 100;
// Give up after POSTGRES_WAIT_TIMEOUT.
let duration = started_at.elapsed();
if duration >= POSTGRES_WAIT_TIMEOUT {
bail!("timed out while waiting for Postgres to start");
}
}
log::info!("PostgreSQL is now running, continuing to configure it");
Ok(())
}

View File

@@ -1,7 +1,9 @@
use std::path::Path;
use std::str::FromStr;
use anyhow::Result;
use log::{info, log_enabled, warn, Level};
use postgres::config::Config;
use postgres::{Client, NoTls};
use serde::Deserialize;
@@ -115,8 +117,8 @@ pub fn handle_roles(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
if existing_roles.iter().any(|r| r.name == op.name) {
let query: String = format!(
"ALTER ROLE {} RENAME TO {}",
op.name.quote(),
new_name.quote()
op.name.pg_quote(),
new_name.pg_quote()
);
warn!("renaming role '{}' to '{}'", op.name, new_name);
@@ -162,7 +164,7 @@ pub fn handle_roles(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
}
if update_role {
let mut query: String = format!("ALTER ROLE {} ", name.quote());
let mut query: String = format!("ALTER ROLE {} ", name.pg_quote());
info_print!(" -> update");
query.push_str(&role.to_pg_options());
@@ -170,7 +172,7 @@ pub fn handle_roles(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
}
} else {
info!("role name: '{}'", &name);
let mut query: String = format!("CREATE ROLE {} ", name.quote());
let mut query: String = format!("CREATE ROLE {} ", name.pg_quote());
info!("role create query: '{}'", &query);
info_print!(" -> create");
@@ -179,7 +181,7 @@ pub fn handle_roles(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
let grant_query = format!(
"GRANT pg_read_all_data, pg_write_all_data TO {}",
name.quote()
name.pg_quote()
);
xact.execute(grant_query.as_str(), &[])?;
info!("role grant query: '{}'", &grant_query);
@@ -215,7 +217,7 @@ pub fn handle_role_deletions(node: &ComputeNode, client: &mut Client) -> Result<
// We do not check either role exists or not,
// Postgres will take care of it for us
if op.action == "delete_role" {
let query: String = format!("DROP ROLE IF EXISTS {}", &op.name.quote());
let query: String = format!("DROP ROLE IF EXISTS {}", &op.name.pg_quote());
warn!("deleting role '{}'", &op.name);
xact.execute(query.as_str(), &[])?;
@@ -230,17 +232,16 @@ pub fn handle_role_deletions(node: &ComputeNode, client: &mut Client) -> Result<
fn reassign_owned_objects(node: &ComputeNode, role_name: &PgIdent) -> Result<()> {
for db in &node.spec.cluster.databases {
if db.owner != *role_name {
let mut connstr = node.connstr.clone();
// database name is always the last and the only component of the path
connstr.set_path(&db.name);
let mut conf = Config::from_str(node.connstr.as_str())?;
conf.dbname(&db.name);
let mut client = Client::connect(connstr.as_str(), NoTls)?;
let mut client = conf.connect(NoTls)?;
// This will reassign all dependent objects to the db owner
let reassign_query = format!(
"REASSIGN OWNED BY {} TO {}",
role_name.quote(),
db.owner.quote()
role_name.pg_quote(),
db.owner.pg_quote()
);
info!(
"reassigning objects owned by '{}' in db '{}' to '{}'",
@@ -249,7 +250,7 @@ fn reassign_owned_objects(node: &ComputeNode, role_name: &PgIdent) -> Result<()>
client.simple_query(&reassign_query)?;
// This now will only drop privileges of the role
let drop_query = format!("DROP OWNED BY {}", role_name.quote());
let drop_query = format!("DROP OWNED BY {}", role_name.pg_quote());
client.simple_query(&drop_query)?;
}
}
@@ -279,7 +280,7 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
// We do not check either DB exists or not,
// Postgres will take care of it for us
"delete_db" => {
let query: String = format!("DROP DATABASE IF EXISTS {}", &op.name.quote());
let query: String = format!("DROP DATABASE IF EXISTS {}", &op.name.pg_quote());
warn!("deleting database '{}'", &op.name);
client.execute(query.as_str(), &[])?;
@@ -291,8 +292,8 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
if existing_dbs.iter().any(|r| r.name == op.name) {
let query: String = format!(
"ALTER DATABASE {} RENAME TO {}",
op.name.quote(),
new_name.quote()
op.name.pg_quote(),
new_name.pg_quote()
);
warn!("renaming database '{}' to '{}'", op.name, new_name);
@@ -320,7 +321,7 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
// XXX: db owner name is returned as quoted string from Postgres,
// when quoting is needed.
let new_owner = if r.owner.starts_with('"') {
db.owner.quote()
db.owner.pg_quote()
} else {
db.owner.clone()
};
@@ -328,15 +329,15 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
if new_owner != r.owner {
let query: String = format!(
"ALTER DATABASE {} OWNER TO {}",
name.quote(),
db.owner.quote()
name.pg_quote(),
db.owner.pg_quote()
);
info_print!(" -> update");
client.execute(query.as_str(), &[])?;
}
} else {
let mut query: String = format!("CREATE DATABASE {} ", name.quote());
let mut query: String = format!("CREATE DATABASE {} ", name.pg_quote());
info_print!(" -> create");
query.push_str(&db.to_pg_options());
@@ -366,7 +367,7 @@ pub fn handle_grants(node: &ComputeNode, client: &mut Client) -> Result<()> {
.cluster
.roles
.iter()
.map(|r| r.name.quote())
.map(|r| r.name.pg_quote())
.collect::<Vec<_>>();
for db in &spec.cluster.databases {
@@ -374,7 +375,7 @@ pub fn handle_grants(node: &ComputeNode, client: &mut Client) -> Result<()> {
let query: String = format!(
"GRANT CREATE ON DATABASE {} TO {}",
dbname.quote(),
dbname.pg_quote(),
roles.join(", ")
);
info!("grant query {}", &query);
@@ -385,12 +386,11 @@ pub fn handle_grants(node: &ComputeNode, client: &mut Client) -> Result<()> {
// Do some per-database access adjustments. We'd better do this at db creation time,
// but CREATE DATABASE isn't transactional. So we cannot create db + do some grants
// atomically.
let mut db_connstr = node.connstr.clone();
for db in &node.spec.cluster.databases {
// database name is always the last and the only component of the path
db_connstr.set_path(&db.name);
let mut conf = Config::from_str(node.connstr.as_str())?;
conf.dbname(&db.name);
let mut db_client = Client::connect(db_connstr.as_str(), NoTls)?;
let mut db_client = conf.connect(NoTls)?;
// This will only change ownership on the schema itself, not the objects
// inside it. Without it owner of the `public` schema will be `cloud_admin`
@@ -419,9 +419,36 @@ pub fn handle_grants(node: &ComputeNode, client: &mut Client) -> Result<()> {
END IF;\n\
END\n\
$$;",
db.owner.quote()
db.owner.pg_quote()
);
db_client.simple_query(&alter_query)?;
// Explicitly grant CREATE ON SCHEMA PUBLIC to the web_access user.
// This is needed because since postgres 15 this privilege is removed by default.
let grant_query = "DO $$\n\
BEGIN\n\
IF EXISTS(\n\
SELECT nspname\n\
FROM pg_catalog.pg_namespace\n\
WHERE nspname = 'public'\n\
) AND\n\
current_setting('server_version_num')::int/10000 >= 15\n\
THEN\n\
IF EXISTS(\n\
SELECT rolname\n\
FROM pg_catalog.pg_roles\n\
WHERE rolname = 'web_access'\n\
)\n\
THEN\n\
GRANT CREATE ON SCHEMA public TO web_access;\n\
END IF;\n\
END IF;\n\
END\n\
$$;"
.to_string();
info!("grant query for db {} : {}", &db.name, &grant_query);
db_client.simple_query(&grant_query)?;
}
Ok(())

View File

@@ -33,9 +33,9 @@ mod pg_helpers_tests {
}
#[test]
fn quote_ident() {
fn ident_pg_quote() {
let ident: PgIdent = PgIdent::from("\"name\";\\n select 1;");
assert_eq!(ident.quote(), "\"\"\"name\"\";\\n select 1;\"");
assert_eq!(ident.pg_quote(), "\"\"\"name\"\";\\n select 1;\"");
}
}

View File

@@ -4,22 +4,24 @@ version = "0.1.0"
edition = "2021"
[dependencies]
clap = "3.0"
comfy-table = "5.0.1"
clap = "4.0"
comfy-table = "6.1"
git-version = "0.3.5"
tar = "0.4.38"
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev = "d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
serde = { version = "1.0", features = ["derive"] }
serde_with = "1.12.0"
serde_with = "2.0"
toml = "0.5"
once_cell = "1.13.0"
regex = "1"
anyhow = "1.0"
thiserror = "1"
nix = "0.23"
nix = "0.25"
reqwest = { version = "0.11", default-features = false, features = ["blocking", "json", "rustls-tls"] }
pageserver = { path = "../pageserver" }
safekeeper = { path = "../safekeeper" }
# Note: Do not directly depend on pageserver or safekeeper; use pageserver_api or safekeeper_api
# instead, so that recompile times are better.
pageserver_api = { path = "../libs/pageserver_api" }
safekeeper_api = { path = "../libs/safekeeper_api" }
utils = { path = "../libs/utils" }
workspace_hack = { version = "0.1", path = "../workspace_hack" }

View File

@@ -6,18 +6,18 @@
//! rely on `neon_local` to set up the environment for each test.
//!
use anyhow::{anyhow, bail, Context, Result};
use clap::{App, AppSettings, Arg, ArgMatches};
use clap::{value_parser, Arg, ArgAction, ArgMatches, Command};
use control_plane::compute::ComputeControlPlane;
use control_plane::local_env::{EtcdBroker, LocalEnv};
use control_plane::safekeeper::SafekeeperNode;
use control_plane::storage::PageServerNode;
use control_plane::{etcd, local_env};
use pageserver::config::defaults::{
use pageserver_api::models::TimelineInfo;
use pageserver_api::{
DEFAULT_HTTP_LISTEN_ADDR as DEFAULT_PAGESERVER_HTTP_ADDR,
DEFAULT_PG_LISTEN_ADDR as DEFAULT_PAGESERVER_PG_ADDR,
};
use pageserver::http::models::TimelineInfo;
use safekeeper::defaults::{
use safekeeper_api::{
DEFAULT_HTTP_LISTEN_PORT as DEFAULT_SAFEKEEPER_HTTP_PORT,
DEFAULT_PG_LISTEN_PORT as DEFAULT_SAFEKEEPER_PG_PORT,
};
@@ -85,212 +85,7 @@ struct TimelineTreeEl {
// * Providing CLI api to the pageserver
// * TODO: export/import to/from usual postgres
fn main() -> Result<()> {
let branch_name_arg = Arg::new("branch-name")
.long("branch-name")
.takes_value(true)
.help("Name of the branch to be created or used as an alias for other services")
.required(false);
let pg_node_arg = Arg::new("node").help("Postgres node name").required(false);
let safekeeper_id_arg = Arg::new("id").help("safekeeper id").required(false);
let tenant_id_arg = Arg::new("tenant-id")
.long("tenant-id")
.help("Tenant id. Represented as a hexadecimal string 32 symbols length")
.takes_value(true)
.required(false);
let timeline_id_arg = Arg::new("timeline-id")
.long("timeline-id")
.help("Timeline id. Represented as a hexadecimal string 32 symbols length")
.takes_value(true)
.required(false);
let pg_version_arg = Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.required(false)
.takes_value(true)
.default_value(DEFAULT_PG_VERSION);
let port_arg = Arg::new("port")
.long("port")
.required(false)
.value_name("port");
let stop_mode_arg = Arg::new("stop-mode")
.short('m')
.takes_value(true)
.possible_values(&["fast", "immediate"])
.help("If 'immediate', don't flush repository data at shutdown")
.required(false)
.value_name("stop-mode");
let pageserver_config_args = Arg::new("pageserver-config-override")
.long("pageserver-config-override")
.takes_value(true)
.number_of_values(1)
.multiple_occurrences(true)
.help("Additional pageserver's configuration options or overrides, refer to pageserver's 'config-override' CLI parameter docs for more")
.required(false);
let lsn_arg = Arg::new("lsn")
.long("lsn")
.help("Specify Lsn on the timeline to start from. By default, end of the timeline would be used.")
.takes_value(true)
.required(false);
let matches = App::new("Neon CLI")
.setting(AppSettings::ArgRequiredElseHelp)
.version(GIT_VERSION)
.subcommand(
App::new("init")
.about("Initialize a new Neon repository")
.arg(pageserver_config_args.clone())
.arg(timeline_id_arg.clone().help("Use a specific timeline id when creating a tenant and its initial timeline"))
.arg(
Arg::new("config")
.long("config")
.required(false)
.value_name("config"),
)
.arg(pg_version_arg.clone())
)
.subcommand(
App::new("timeline")
.about("Manage timelines")
.subcommand(App::new("list")
.about("List all timelines, available to this pageserver")
.arg(tenant_id_arg.clone()))
.subcommand(App::new("branch")
.about("Create a new timeline, using another timeline as a base, copying its data")
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone())
.arg(Arg::new("ancestor-branch-name").long("ancestor-branch-name").takes_value(true)
.help("Use last Lsn of another timeline (and its data) as base when creating the new timeline. The timeline gets resolved by its branch name.").required(false))
.arg(Arg::new("ancestor-start-lsn").long("ancestor-start-lsn").takes_value(true)
.help("When using another timeline as base, use a specific Lsn in it instead of the latest one").required(false)))
.subcommand(App::new("create")
.about("Create a new blank timeline")
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone())
.arg(pg_version_arg.clone())
)
.subcommand(App::new("import")
.about("Import timeline from basebackup directory")
.arg(tenant_id_arg.clone())
.arg(timeline_id_arg.clone())
.arg(Arg::new("node-name").long("node-name").takes_value(true)
.help("Name to assign to the imported timeline"))
.arg(Arg::new("base-tarfile").long("base-tarfile").takes_value(true)
.help("Basebackup tarfile to import"))
.arg(Arg::new("base-lsn").long("base-lsn").takes_value(true)
.help("Lsn the basebackup starts at"))
.arg(Arg::new("wal-tarfile").long("wal-tarfile").takes_value(true)
.help("Wal to add after base"))
.arg(Arg::new("end-lsn").long("end-lsn").takes_value(true)
.help("Lsn the basebackup ends at"))
.arg(pg_version_arg.clone())
)
).subcommand(
App::new("tenant")
.setting(AppSettings::ArgRequiredElseHelp)
.about("Manage tenants")
.subcommand(App::new("list"))
.subcommand(App::new("create")
.arg(tenant_id_arg.clone())
.arg(timeline_id_arg.clone().help("Use a specific timeline id when creating a tenant and its initial timeline"))
.arg(Arg::new("config").short('c').takes_value(true).multiple_occurrences(true).required(false))
.arg(pg_version_arg.clone())
)
.subcommand(App::new("config")
.arg(tenant_id_arg.clone())
.arg(Arg::new("config").short('c').takes_value(true).multiple_occurrences(true).required(false))
)
)
.subcommand(
App::new("pageserver")
.setting(AppSettings::ArgRequiredElseHelp)
.about("Manage pageserver")
.subcommand(App::new("status"))
.subcommand(App::new("start").about("Start local pageserver").arg(pageserver_config_args.clone()))
.subcommand(App::new("stop").about("Stop local pageserver")
.arg(stop_mode_arg.clone()))
.subcommand(App::new("restart").about("Restart local pageserver").arg(pageserver_config_args.clone()))
)
.subcommand(
App::new("safekeeper")
.setting(AppSettings::ArgRequiredElseHelp)
.about("Manage safekeepers")
.subcommand(App::new("start")
.about("Start local safekeeper")
.arg(safekeeper_id_arg.clone())
)
.subcommand(App::new("stop")
.about("Stop local safekeeper")
.arg(safekeeper_id_arg.clone())
.arg(stop_mode_arg.clone())
)
.subcommand(App::new("restart")
.about("Restart local safekeeper")
.arg(safekeeper_id_arg.clone())
.arg(stop_mode_arg.clone())
)
)
.subcommand(
App::new("pg")
.setting(AppSettings::ArgRequiredElseHelp)
.about("Manage postgres instances")
.subcommand(App::new("list").arg(tenant_id_arg.clone()))
.subcommand(App::new("create")
.about("Create a postgres compute node")
.arg(pg_node_arg.clone())
.arg(branch_name_arg.clone())
.arg(tenant_id_arg.clone())
.arg(lsn_arg.clone())
.arg(port_arg.clone())
.arg(
Arg::new("config-only")
.help("Don't do basebackup, create compute node with only config files")
.long("config-only")
.required(false))
.arg(pg_version_arg.clone())
)
.subcommand(App::new("start")
.about("Start a postgres compute node.\n This command actually creates new node from scratch, but preserves existing config files")
.arg(pg_node_arg.clone())
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone())
.arg(timeline_id_arg.clone())
.arg(lsn_arg.clone())
.arg(port_arg.clone())
.arg(pg_version_arg.clone())
)
.subcommand(
App::new("stop")
.arg(pg_node_arg.clone())
.arg(tenant_id_arg.clone())
.arg(
Arg::new("destroy")
.help("Also delete data directory (now optional, should be default in future)")
.long("destroy")
.required(false)
)
)
)
.subcommand(
App::new("start")
.about("Start page server and safekeepers")
.arg(pageserver_config_args)
)
.subcommand(
App::new("stop")
.about("Stop page server and safekeepers")
.arg(stop_mode_arg.clone())
)
.get_matches();
let matches = cli().get_matches();
let (sub_name, sub_args) = match matches.subcommand() {
Some(subcommand_data) => subcommand_data,
@@ -358,9 +153,7 @@ fn print_timelines_tree(
// Memorize all direct children of each timeline.
for timeline in timelines.iter() {
if let Some(ancestor_timeline_id) =
timeline.local.as_ref().and_then(|l| l.ancestor_timeline_id)
{
if let Some(ancestor_timeline_id) = timeline.ancestor_timeline_id {
timelines_hash
.get_mut(&ancestor_timeline_id)
.context("missing timeline info in the HashMap")?
@@ -371,13 +164,7 @@ fn print_timelines_tree(
for timeline in timelines_hash.values() {
// Start with root local timelines (no ancestors) first.
if timeline
.info
.local
.as_ref()
.and_then(|l| l.ancestor_timeline_id)
.is_none()
{
if timeline.info.ancestor_timeline_id.is_none() {
print_timeline(0, &Vec::from([true]), timeline, &timelines_hash)?;
}
}
@@ -394,17 +181,8 @@ fn print_timeline(
timeline: &TimelineTreeEl,
timelines: &HashMap<TimelineId, TimelineTreeEl>,
) -> Result<()> {
let local_remote = match (timeline.info.local.as_ref(), timeline.info.remote.as_ref()) {
(None, None) => unreachable!("in this case no info for a timeline is found"),
(None, Some(_)) => "(R)",
(Some(_), None) => "(L)",
(Some(_), Some(_)) => "(L+R)",
};
// Draw main padding
print!("{} ", local_remote);
if nesting_level > 0 {
let ancestor_lsn = match timeline.info.local.as_ref().and_then(|i| i.ancestor_lsn) {
let ancestor_lsn = match timeline.info.ancestor_lsn {
Some(lsn) => lsn.to_string(),
None => "Unknown Lsn".to_string(),
};
@@ -492,16 +270,16 @@ fn get_tenant_id(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::R
fn parse_tenant_id(sub_match: &ArgMatches) -> anyhow::Result<Option<TenantId>> {
sub_match
.value_of("tenant-id")
.map(TenantId::from_str)
.get_one::<String>("tenant-id")
.map(|tenant_id| TenantId::from_str(tenant_id))
.transpose()
.context("Failed to parse tenant id from the argument string")
}
fn parse_timeline_id(sub_match: &ArgMatches) -> anyhow::Result<Option<TimelineId>> {
sub_match
.value_of("timeline-id")
.map(TimelineId::from_str)
.get_one::<String>("timeline-id")
.map(|timeline_id| TimelineId::from_str(timeline_id))
.transpose()
.context("Failed to parse timeline id from the argument string")
}
@@ -510,19 +288,22 @@ fn handle_init(init_match: &ArgMatches) -> anyhow::Result<LocalEnv> {
let initial_timeline_id_arg = parse_timeline_id(init_match)?;
// Create config file
let toml_file: String = if let Some(config_path) = init_match.value_of("config") {
let toml_file: String = if let Some(config_path) = init_match.get_one::<PathBuf>("config") {
// load and parse the file
std::fs::read_to_string(std::path::Path::new(config_path))
.with_context(|| format!("Could not read configuration file '{config_path}'"))?
std::fs::read_to_string(config_path).with_context(|| {
format!(
"Could not read configuration file '{}'",
config_path.display()
)
})?
} else {
// Built-in default config
default_conf(&EtcdBroker::locate_etcd()?)
};
let pg_version = init_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.get_one::<u32>("pg-version")
.copied()
.context("Failed to parse postgres version from the argument string")?;
let mut env =
@@ -558,9 +339,10 @@ fn handle_init(init_match: &ArgMatches) -> anyhow::Result<LocalEnv> {
fn pageserver_config_overrides(init_match: &ArgMatches) -> Vec<&str> {
init_match
.values_of("pageserver-config-override")
.get_many::<String>("pageserver-config-override")
.into_iter()
.flatten()
.map(|s| s.as_str())
.collect()
}
@@ -575,7 +357,7 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
Some(("create", create_match)) => {
let initial_tenant_id = parse_tenant_id(create_match)?;
let tenant_conf: HashMap<_, _> = create_match
.values_of("config")
.get_many::<String>("config")
.map(|vals| vals.flat_map(|c| c.split_once(':')).collect())
.unwrap_or_default();
let new_tenant_id = pageserver.tenant_create(initial_tenant_id, tenant_conf)?;
@@ -584,9 +366,8 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
// Create an initial timeline for the new tenant
let new_timeline_id = parse_timeline_id(create_match)?;
let pg_version = create_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.get_one::<u32>("pg-version")
.copied()
.context("Failed to parse postgres version from the argument string")?;
let timeline_info = pageserver.timeline_create(
@@ -597,10 +378,7 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
Some(pg_version),
)?;
let new_timeline_id = timeline_info.timeline_id;
let last_record_lsn = timeline_info
.local
.context(format!("Failed to get last record LSN: no local timeline info for timeline {new_timeline_id}"))?
.last_record_lsn;
let last_record_lsn = timeline_info.last_record_lsn;
env.register_branch_mapping(
DEFAULT_BRANCH_NAME.to_string(),
@@ -615,7 +393,7 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
Some(("config", create_match)) => {
let tenant_id = get_tenant_id(create_match, env)?;
let tenant_conf: HashMap<_, _> = create_match
.values_of("config")
.get_many::<String>("config")
.map(|vals| vals.flat_map(|c| c.split_once(':')).collect())
.unwrap_or_default();
@@ -642,23 +420,19 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
Some(("create", create_match)) => {
let tenant_id = get_tenant_id(create_match, env)?;
let new_branch_name = create_match
.value_of("branch-name")
.get_one::<String>("branch-name")
.ok_or_else(|| anyhow!("No branch name provided"))?;
let pg_version = create_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.get_one::<u32>("pg-version")
.copied()
.context("Failed to parse postgres version from the argument string")?;
let timeline_info =
pageserver.timeline_create(tenant_id, None, None, None, Some(pg_version))?;
let new_timeline_id = timeline_info.timeline_id;
let last_record_lsn = timeline_info
.local
.expect("no local timeline info")
.last_record_lsn;
let last_record_lsn = timeline_info.last_record_lsn;
env.register_branch_mapping(new_branch_name.to_string(), tenant_id, new_timeline_id)?;
println!(
@@ -670,35 +444,32 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
let tenant_id = get_tenant_id(import_match, env)?;
let timeline_id = parse_timeline_id(import_match)?.expect("No timeline id provided");
let name = import_match
.value_of("node-name")
.get_one::<String>("node-name")
.ok_or_else(|| anyhow!("No node name provided"))?;
// Parse base inputs
let base_tarfile = import_match
.value_of("base-tarfile")
.map(|s| PathBuf::from_str(s).unwrap())
.ok_or_else(|| anyhow!("No base-tarfile provided"))?;
.get_one::<PathBuf>("base-tarfile")
.ok_or_else(|| anyhow!("No base-tarfile provided"))?
.to_owned();
let base_lsn = Lsn::from_str(
import_match
.value_of("base-lsn")
.get_one::<String>("base-lsn")
.ok_or_else(|| anyhow!("No base-lsn provided"))?,
)?;
let base = (base_lsn, base_tarfile);
// Parse pg_wal inputs
let wal_tarfile = import_match
.value_of("wal-tarfile")
.map(|s| PathBuf::from_str(s).unwrap());
let wal_tarfile = import_match.get_one::<PathBuf>("wal-tarfile").cloned();
let end_lsn = import_match
.value_of("end-lsn")
.get_one::<String>("end-lsn")
.map(|s| Lsn::from_str(s).unwrap());
// TODO validate both or none are provided
let pg_wal = end_lsn.zip(wal_tarfile);
let pg_version = import_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.get_one::<u32>("pg-version")
.copied()
.context("Failed to parse postgres version from the argument string")?;
let mut cplane = ComputeControlPlane::load(env.clone())?;
@@ -713,10 +484,11 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
Some(("branch", branch_match)) => {
let tenant_id = get_tenant_id(branch_match, env)?;
let new_branch_name = branch_match
.value_of("branch-name")
.get_one::<String>("branch-name")
.ok_or_else(|| anyhow!("No branch name provided"))?;
let ancestor_branch_name = branch_match
.value_of("ancestor-branch-name")
.get_one::<String>("ancestor-branch-name")
.map(|s| s.as_str())
.unwrap_or(DEFAULT_BRANCH_NAME);
let ancestor_timeline_id = env
.get_branch_timeline_id(ancestor_branch_name, tenant_id)
@@ -725,8 +497,8 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
})?;
let start_lsn = branch_match
.value_of("ancestor-start-lsn")
.map(Lsn::from_str)
.get_one::<String>("ancestor-start-lsn")
.map(|lsn_str| Lsn::from_str(lsn_str))
.transpose()
.context("Failed to parse ancestor start Lsn from the request")?;
let timeline_info = pageserver.timeline_create(
@@ -738,10 +510,7 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
)?;
let new_timeline_id = timeline_info.timeline_id;
let last_record_lsn = timeline_info
.local
.expect("no local timeline info")
.last_record_lsn;
let last_record_lsn = timeline_info.last_record_lsn;
env.register_branch_mapping(new_branch_name.to_string(), tenant_id, new_timeline_id)?;
@@ -801,7 +570,7 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
// Use the LSN at the end of the timeline.
timeline_infos
.get(&node.timeline_id)
.and_then(|bi| bi.local.as_ref().map(|l| l.last_record_lsn.to_string()))
.map(|bi| bi.last_record_lsn.to_string())
.unwrap_or_else(|| "?".to_string())
}
Some(lsn) => {
@@ -830,45 +599,39 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
}
"create" => {
let branch_name = sub_args
.value_of("branch-name")
.get_one::<String>("branch-name")
.map(|s| s.as_str())
.unwrap_or(DEFAULT_BRANCH_NAME);
let node_name = sub_args
.value_of("node")
.map(ToString::to_string)
.unwrap_or_else(|| format!("{}_node", branch_name));
.get_one::<String>("node")
.map(|node_name| node_name.to_string())
.unwrap_or_else(|| format!("{branch_name}_node"));
let lsn = sub_args
.value_of("lsn")
.map(Lsn::from_str)
.get_one::<String>("lsn")
.map(|lsn_str| Lsn::from_str(lsn_str))
.transpose()
.context("Failed to parse Lsn from the request")?;
let timeline_id = env
.get_branch_timeline_id(branch_name, tenant_id)
.ok_or_else(|| anyhow!("Found no timeline id for branch name '{}'", branch_name))?;
.ok_or_else(|| anyhow!("Found no timeline id for branch name '{branch_name}'"))?;
let port: Option<u16> = match sub_args.value_of("port") {
Some(p) => Some(p.parse()?),
None => None,
};
let port: Option<u16> = sub_args.get_one::<u16>("port").copied();
let pg_version = sub_args
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.get_one::<u32>("pg-version")
.copied()
.context("Failed to parse postgres version from the argument string")?;
cplane.new_node(tenant_id, &node_name, timeline_id, lsn, port, pg_version)?;
}
"start" => {
let port: Option<u16> = match sub_args.value_of("port") {
Some(p) => Some(p.parse()?),
None => None,
};
let port: Option<u16> = sub_args.get_one::<u16>("port").copied();
let node_name = sub_args
.value_of("node")
.get_one::<String>("node")
.ok_or_else(|| anyhow!("No node name was provided to start"))?;
let node = cplane.nodes.get(&(tenant_id, node_name.to_owned()));
let node = cplane.nodes.get(&(tenant_id, node_name.to_string()));
let auth_token = if matches!(env.pageserver.auth_type, AuthType::NeonJWT) {
let claims = Claims::new(Some(tenant_id), Scope::Tenant);
@@ -879,36 +642,33 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
};
if let Some(node) = node {
println!("Starting existing postgres {}...", node_name);
println!("Starting existing postgres {node_name}...");
node.start(&auth_token)?;
} else {
let branch_name = sub_args
.value_of("branch-name")
.get_one::<String>("branch-name")
.map(|s| s.as_str())
.unwrap_or(DEFAULT_BRANCH_NAME);
let timeline_id = env
.get_branch_timeline_id(branch_name, tenant_id)
.ok_or_else(|| {
anyhow!("Found no timeline id for branch name '{}'", branch_name)
anyhow!("Found no timeline id for branch name '{branch_name}'")
})?;
let lsn = sub_args
.value_of("lsn")
.map(Lsn::from_str)
.get_one::<String>("lsn")
.map(|lsn_str| Lsn::from_str(lsn_str))
.transpose()
.context("Failed to parse Lsn from the request")?;
let pg_version = sub_args
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
.get_one::<u32>("pg-version")
.copied()
.context("Failed to `pg-version` from the argument string")?;
// when used with custom port this results in non obvious behaviour
// port is remembered from first start command, i e
// start --port X
// stop
// start <-- will also use port X even without explicit port argument
println!(
"Starting new postgres (v{}) {} on timeline {} ...",
pg_version, node_name, timeline_id
);
println!("Starting new postgres (v{pg_version}) {node_name} on timeline {timeline_id} ...");
let node =
cplane.new_node(tenant_id, node_name, timeline_id, lsn, port, pg_version)?;
@@ -917,18 +677,18 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
}
"stop" => {
let node_name = sub_args
.value_of("node")
.get_one::<String>("node")
.ok_or_else(|| anyhow!("No node name was provided to stop"))?;
let destroy = sub_args.is_present("destroy");
let destroy = sub_args.get_flag("destroy");
let node = cplane
.nodes
.get(&(tenant_id, node_name.to_owned()))
.with_context(|| format!("postgres {} is not found", node_name))?;
.get(&(tenant_id, node_name.to_string()))
.with_context(|| format!("postgres {node_name} is not found"))?;
node.stop(destroy)?;
}
_ => bail!("Unexpected pg subcommand '{}'", sub_name),
_ => bail!("Unexpected pg subcommand '{sub_name}'"),
}
Ok(())
@@ -946,7 +706,10 @@ fn handle_pageserver(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Resul
}
Some(("stop", stop_match)) => {
let immediate = stop_match.value_of("stop-mode") == Some("immediate");
let immediate = stop_match
.get_one::<String>("stop-mode")
.map(|s| s.as_str())
== Some("immediate");
if let Err(e) = pageserver.stop(immediate) {
eprintln!("pageserver stop failed: {}", e);
@@ -996,7 +759,7 @@ fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Resul
};
// All the commands take an optional safekeeper name argument
let sk_id = if let Some(id_str) = sub_args.value_of("id") {
let sk_id = if let Some(id_str) = sub_args.get_one::<String>("id") {
NodeId(id_str.parse().context("while parsing safekeeper id")?)
} else {
DEFAULT_SAFEKEEPER_ID
@@ -1012,7 +775,8 @@ fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Resul
}
"stop" => {
let immediate = sub_args.value_of("stop-mode") == Some("immediate");
let immediate =
sub_args.get_one::<String>("stop-mode").map(|s| s.as_str()) == Some("immediate");
if let Err(e) = safekeeper.stop(immediate) {
eprintln!("safekeeper stop failed: {}", e);
@@ -1021,7 +785,8 @@ fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Resul
}
"restart" => {
let immediate = sub_args.value_of("stop-mode") == Some("immediate");
let immediate =
sub_args.get_one::<String>("stop-mode").map(|s| s.as_str()) == Some("immediate");
if let Err(e) = safekeeper.stop(immediate) {
eprintln!("safekeeper stop failed: {}", e);
@@ -1065,7 +830,8 @@ fn handle_start_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow
}
fn handle_stop_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
let immediate = sub_match.value_of("stop-mode") == Some("immediate");
let immediate =
sub_match.get_one::<String>("stop-mode").map(|s| s.as_str()) == Some("immediate");
let pageserver = PageServerNode::from_env(env);
@@ -1098,3 +864,219 @@ fn try_stop_etcd_process(env: &local_env::LocalEnv) {
eprintln!("etcd stop failed: {e}");
}
}
fn cli() -> Command {
let branch_name_arg = Arg::new("branch-name")
.long("branch-name")
.help("Name of the branch to be created or used as an alias for other services")
.required(false);
let pg_node_arg = Arg::new("node").help("Postgres node name").required(false);
let safekeeper_id_arg = Arg::new("id").help("safekeeper id").required(false);
let tenant_id_arg = Arg::new("tenant-id")
.long("tenant-id")
.help("Tenant id. Represented as a hexadecimal string 32 symbols length")
.required(false);
let timeline_id_arg = Arg::new("timeline-id")
.long("timeline-id")
.help("Timeline id. Represented as a hexadecimal string 32 symbols length")
.required(false);
let pg_version_arg = Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.required(false)
.value_parser(value_parser!(u32))
.default_value(DEFAULT_PG_VERSION);
let port_arg = Arg::new("port")
.long("port")
.required(false)
.value_parser(value_parser!(u16))
.value_name("port");
let stop_mode_arg = Arg::new("stop-mode")
.short('m')
.value_parser(["fast", "immediate"])
.help("If 'immediate', don't flush repository data at shutdown")
.required(false)
.value_name("stop-mode");
let pageserver_config_args = Arg::new("pageserver-config-override")
.long("pageserver-config-override")
.num_args(1)
.action(ArgAction::Append)
.help("Additional pageserver's configuration options or overrides, refer to pageserver's 'config-override' CLI parameter docs for more")
.required(false);
let lsn_arg = Arg::new("lsn")
.long("lsn")
.help("Specify Lsn on the timeline to start from. By default, end of the timeline would be used.")
.required(false);
Command::new("Neon CLI")
.arg_required_else_help(true)
.version(GIT_VERSION)
.subcommand(
Command::new("init")
.about("Initialize a new Neon repository")
.arg(pageserver_config_args.clone())
.arg(timeline_id_arg.clone().help("Use a specific timeline id when creating a tenant and its initial timeline"))
.arg(
Arg::new("config")
.long("config")
.required(false)
.value_parser(value_parser!(PathBuf))
.value_name("config"),
)
.arg(pg_version_arg.clone())
)
.subcommand(
Command::new("timeline")
.about("Manage timelines")
.subcommand(Command::new("list")
.about("List all timelines, available to this pageserver")
.arg(tenant_id_arg.clone()))
.subcommand(Command::new("branch")
.about("Create a new timeline, using another timeline as a base, copying its data")
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone())
.arg(Arg::new("ancestor-branch-name").long("ancestor-branch-name")
.help("Use last Lsn of another timeline (and its data) as base when creating the new timeline. The timeline gets resolved by its branch name.").required(false))
.arg(Arg::new("ancestor-start-lsn").long("ancestor-start-lsn")
.help("When using another timeline as base, use a specific Lsn in it instead of the latest one").required(false)))
.subcommand(Command::new("create")
.about("Create a new blank timeline")
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone())
.arg(pg_version_arg.clone())
)
.subcommand(Command::new("import")
.about("Import timeline from basebackup directory")
.arg(tenant_id_arg.clone())
.arg(timeline_id_arg.clone())
.arg(Arg::new("node-name").long("node-name")
.help("Name to assign to the imported timeline"))
.arg(Arg::new("base-tarfile")
.long("base-tarfile")
.value_parser(value_parser!(PathBuf))
.help("Basebackup tarfile to import")
)
.arg(Arg::new("base-lsn").long("base-lsn")
.help("Lsn the basebackup starts at"))
.arg(Arg::new("wal-tarfile")
.long("wal-tarfile")
.value_parser(value_parser!(PathBuf))
.help("Wal to add after base")
)
.arg(Arg::new("end-lsn").long("end-lsn")
.help("Lsn the basebackup ends at"))
.arg(pg_version_arg.clone())
)
).subcommand(
Command::new("tenant")
.arg_required_else_help(true)
.about("Manage tenants")
.subcommand(Command::new("list"))
.subcommand(Command::new("create")
.arg(tenant_id_arg.clone())
.arg(timeline_id_arg.clone().help("Use a specific timeline id when creating a tenant and its initial timeline"))
.arg(Arg::new("config").short('c').num_args(1).action(ArgAction::Append).required(false))
.arg(pg_version_arg.clone())
)
.subcommand(Command::new("config")
.arg(tenant_id_arg.clone())
.arg(Arg::new("config").short('c').num_args(1).action(ArgAction::Append).required(false))
)
)
.subcommand(
Command::new("pageserver")
.arg_required_else_help(true)
.about("Manage pageserver")
.subcommand(Command::new("status"))
.subcommand(Command::new("start").about("Start local pageserver").arg(pageserver_config_args.clone()))
.subcommand(Command::new("stop").about("Stop local pageserver")
.arg(stop_mode_arg.clone()))
.subcommand(Command::new("restart").about("Restart local pageserver").arg(pageserver_config_args.clone()))
)
.subcommand(
Command::new("safekeeper")
.arg_required_else_help(true)
.about("Manage safekeepers")
.subcommand(Command::new("start")
.about("Start local safekeeper")
.arg(safekeeper_id_arg.clone())
)
.subcommand(Command::new("stop")
.about("Stop local safekeeper")
.arg(safekeeper_id_arg.clone())
.arg(stop_mode_arg.clone())
)
.subcommand(Command::new("restart")
.about("Restart local safekeeper")
.arg(safekeeper_id_arg)
.arg(stop_mode_arg.clone())
)
)
.subcommand(
Command::new("pg")
.arg_required_else_help(true)
.about("Manage postgres instances")
.subcommand(Command::new("list").arg(tenant_id_arg.clone()))
.subcommand(Command::new("create")
.about("Create a postgres compute node")
.arg(pg_node_arg.clone())
.arg(branch_name_arg.clone())
.arg(tenant_id_arg.clone())
.arg(lsn_arg.clone())
.arg(port_arg.clone())
.arg(
Arg::new("config-only")
.help("Don't do basebackup, create compute node with only config files")
.long("config-only")
.required(false))
.arg(pg_version_arg.clone())
)
.subcommand(Command::new("start")
.about("Start a postgres compute node.\n This command actually creates new node from scratch, but preserves existing config files")
.arg(pg_node_arg.clone())
.arg(tenant_id_arg.clone())
.arg(branch_name_arg)
.arg(timeline_id_arg)
.arg(lsn_arg)
.arg(port_arg)
.arg(pg_version_arg)
)
.subcommand(
Command::new("stop")
.arg(pg_node_arg)
.arg(tenant_id_arg)
.arg(
Arg::new("destroy")
.help("Also delete data directory (now optional, should be default in future)")
.long("destroy")
.action(ArgAction::SetTrue)
.required(false)
)
)
)
.subcommand(
Command::new("start")
.about("Start page server and safekeepers")
.arg(pageserver_config_args)
)
.subcommand(
Command::new("stop")
.about("Stop page server and safekeepers")
.arg(stop_mode_arg)
)
}
#[test]
fn verify_cli() {
cli().debug_assert();
}

View File

@@ -183,18 +183,18 @@ impl PostgresNode {
}
fn sync_safekeepers(&self, auth_token: &Option<String>, pg_version: u32) -> Result<Lsn> {
let pg_path = self.env.pg_bin_dir(pg_version).join("postgres");
let pg_path = self.env.pg_bin_dir(pg_version)?.join("postgres");
let mut cmd = Command::new(&pg_path);
cmd.arg("--sync-safekeepers")
.env_clear()
.env(
"LD_LIBRARY_PATH",
self.env.pg_lib_dir(pg_version).to_str().unwrap(),
self.env.pg_lib_dir(pg_version)?.to_str().unwrap(),
)
.env(
"DYLD_LIBRARY_PATH",
self.env.pg_lib_dir(pg_version).to_str().unwrap(),
self.env.pg_lib_dir(pg_version)?.to_str().unwrap(),
)
.env("PGDATA", self.pgdata().to_str().unwrap())
.stdout(Stdio::piped())
@@ -282,9 +282,7 @@ impl PostgresNode {
fn setup_pg_conf(&self, auth_type: AuthType) -> Result<()> {
let mut conf = PostgresConf::new();
conf.append("max_wal_senders", "10");
// wal_log_hints is mandatory when running against pageserver (see gh issue#192)
// TODO: is it possible to check wal_log_hints at pageserver side via XLOG_PARAMETER_CHANGE?
conf.append("wal_log_hints", "on");
conf.append("wal_log_hints", "off");
conf.append("max_replication_slots", "10");
conf.append("hot_standby", "on");
conf.append("shared_buffers", "1MB");
@@ -422,7 +420,7 @@ impl PostgresNode {
}
fn pg_ctl(&self, args: &[&str], auth_token: &Option<String>) -> Result<()> {
let pg_ctl_path = self.env.pg_bin_dir(self.pg_version).join("pg_ctl");
let pg_ctl_path = self.env.pg_bin_dir(self.pg_version)?.join("pg_ctl");
let mut cmd = Command::new(pg_ctl_path);
cmd.args(
[
@@ -440,11 +438,11 @@ impl PostgresNode {
.env_clear()
.env(
"LD_LIBRARY_PATH",
self.env.pg_lib_dir(self.pg_version).to_str().unwrap(),
self.env.pg_lib_dir(self.pg_version)?.to_str().unwrap(),
)
.env(
"DYLD_LIBRARY_PATH",
self.env.pg_lib_dir(self.pg_version).to_str().unwrap(),
self.env.pg_lib_dir(self.pg_version)?.to_str().unwrap(),
);
if let Some(token) = auth_token {
cmd.env("ZENITH_AUTH_TOKEN", token);

View File

@@ -52,6 +52,10 @@ pub fn start_etcd_process(env: &local_env::LocalEnv) -> anyhow::Result<()> {
// size smaller. Our test etcd clusters are very small.
// See https://github.com/etcd-io/etcd/issues/7910
"--quota-backend-bytes=100000000".to_string(),
// etcd doesn't compact (vacuum) with default settings,
// enable it to prevent space exhaustion.
"--auto-compaction-mode=revision".to_string(),
"--auto-compaction-retention=1".to_string(),
])
.stdout(Stdio::from(etcd_stdout_file))
.stderr(Stdio::from(etcd_stderr_file))

View File

@@ -201,28 +201,28 @@ impl LocalEnv {
self.pg_distrib_dir.clone()
}
pub fn pg_distrib_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_distrib_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
let path = self.pg_distrib_dir.clone();
match pg_version {
14 => path.join(format!("v{pg_version}")),
15 => path.join(format!("v{pg_version}")),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(path.join(format!("v{pg_version}"))),
15 => Ok(path.join(format!("v{pg_version}"))),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_bin_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_bin_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("bin"),
15 => self.pg_distrib_dir(pg_version).join("bin"),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(self.pg_distrib_dir(pg_version)?.join("bin")),
15 => Ok(self.pg_distrib_dir(pg_version)?.join("bin")),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_lib_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_lib_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("lib"),
15 => self.pg_distrib_dir(pg_version).join("lib"),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(self.pg_distrib_dir(pg_version)?.join("lib")),
15 => Ok(self.pg_distrib_dir(pg_version)?.join("lib")),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
@@ -285,7 +285,7 @@ impl LocalEnv {
branch_name: &str,
tenant_id: TenantId,
) -> Option<TimelineId> {
dbg!(&self.branch_name_mappings)
self.branch_name_mappings
.get(branch_name)?
.iter()
.find(|(mapped_tenant_id, _)| mapped_tenant_id == &tenant_id)
@@ -422,10 +422,10 @@ impl LocalEnv {
"directory '{}' already exists. Perhaps already initialized?",
base_path.display()
);
if !self.pg_bin_dir(pg_version).join("postgres").exists() {
if !self.pg_bin_dir(pg_version)?.join("postgres").exists() {
bail!(
"Can't find postgres binary at {}",
self.pg_bin_dir(pg_version).display()
self.pg_bin_dir(pg_version)?.display()
);
}
for binary in ["pageserver", "safekeeper"] {

View File

@@ -12,13 +12,8 @@ use nix::unistd::Pid;
use postgres::Config;
use reqwest::blocking::{Client, RequestBuilder, Response};
use reqwest::{IntoUrl, Method};
use safekeeper::http::models::TimelineCreateRequest;
use thiserror::Error;
use utils::{
connstring::connection_address,
http::error::HttpErrorBody,
id::{NodeId, TenantId, TimelineId},
};
use utils::{connstring::connection_address, http::error::HttpErrorBody, id::NodeId};
use crate::local_env::{LocalEnv, SafekeeperConf};
use crate::storage::PageServerNode;
@@ -128,7 +123,6 @@ impl SafekeeperNode {
.args(&["--id", self.id.to_string().as_ref()])
.args(&["--listen-pg", &listen_pg])
.args(&["--listen-http", &listen_http])
.args(&["--recall", "1 second"])
.arg("--daemonize"),
);
if !self.conf.sync {
@@ -281,24 +275,4 @@ impl SafekeeperNode {
.error_from_body()?;
Ok(())
}
pub fn timeline_create(
&self,
tenant_id: TenantId,
timeline_id: TimelineId,
peer_ids: Vec<NodeId>,
) -> Result<()> {
Ok(self
.http_request(
Method::POST,
format!("{}/tenant/{}/timeline", self.http_base_url, tenant_id),
)
.json(&TimelineCreateRequest {
timeline_id,
peer_ids,
})
.send()?
.error_from_body()?
.json()?)
}
}

View File

@@ -11,7 +11,7 @@ use anyhow::{bail, Context};
use nix::errno::Errno;
use nix::sys::signal::{kill, Signal};
use nix::unistd::Pid;
use pageserver::http::models::{
use pageserver_api::models::{
TenantConfigRequest, TenantCreateRequest, TenantInfo, TimelineCreateRequest, TimelineInfo,
};
use postgres::{Config, NoTls};

View File

@@ -0,0 +1,48 @@
#!/bin/bash
set -eux
PG_VERSION=${PG_VERSION:-14}
SPEC_FILE_ORG=/var/db/postgres/specs/spec.json
SPEC_FILE=/tmp/spec.json
echo "Waiting pageserver become ready."
while ! nc -z pageserver 6400; do
sleep 1;
done
echo "Page server is ready."
echo "Create a tenant and timeline"
PARAMS=(
-sb
-X POST
-H "Content-Type: application/json"
-d "{}"
http://pageserver:9898/v1/tenant/
)
tenant_id=$(curl "${PARAMS[@]}" | sed 's/"//g')
PARAMS=(
-sb
-X POST
-H "Content-Type: application/json"
-d "{\"tenant_id\":\"${tenant_id}\", \"pg_version\": ${PG_VERSION}}"
"http://pageserver:9898/v1/tenant/${tenant_id}/timeline/"
)
result=$(curl "${PARAMS[@]}")
echo $result | jq .
echo "Overwrite tenant id and timeline id in spec file"
tenant_id=$(echo ${result} | jq -r .tenant_id)
timeline_id=$(echo ${result} | jq -r .timeline_id)
sed "s/TENANT_ID/${tenant_id}/" ${SPEC_FILE_ORG} > ${SPEC_FILE}
sed -i "s/TIMELINE_ID/${timeline_id}/" ${SPEC_FILE}
cat ${SPEC_FILE}
echo "Start compute node"
/usr/local/bin/compute_ctl --pgdata /var/db/postgres/compute \
-C "postgresql://cloud_admin@localhost:55433/postgres" \
-b /usr/local/bin/postgres \
-S ${SPEC_FILE}

View File

@@ -0,0 +1,141 @@
{
"format_version": 1.0,
"timestamp": "2022-10-12T18:00:00.000Z",
"operation_uuid": "0f657b36-4b0f-4a2d-9c2e-1dcd615e7d8c",
"cluster": {
"cluster_id": "docker_compose",
"name": "docker_compose_test",
"state": "restarted",
"roles": [
{
"name": "cloud_admin",
"encrypted_password": "b093c0d3b281ba6da1eacc608620abd8",
"options": null
}
],
"databases": [
],
"settings": [
{
"name": "fsync",
"value": "off",
"vartype": "bool"
},
{
"name": "wal_level",
"value": "replica",
"vartype": "enum"
},
{
"name": "hot_standby",
"value": "on",
"vartype": "bool"
},
{
"name": "wal_log_hints",
"value": "on",
"vartype": "bool"
},
{
"name": "log_connections",
"value": "on",
"vartype": "bool"
},
{
"name": "port",
"value": "55433",
"vartype": "integer"
},
{
"name": "shared_buffers",
"value": "1MB",
"vartype": "string"
},
{
"name": "max_connections",
"value": "100",
"vartype": "integer"
},
{
"name": "listen_addresses",
"value": "0.0.0.0",
"vartype": "string"
},
{
"name": "max_wal_senders",
"value": "10",
"vartype": "integer"
},
{
"name": "max_replication_slots",
"value": "10",
"vartype": "integer"
},
{
"name": "wal_sender_timeout",
"value": "5s",
"vartype": "string"
},
{
"name": "wal_keep_size",
"value": "0",
"vartype": "integer"
},
{
"name": "password_encryption",
"value": "md5",
"vartype": "enum"
},
{
"name": "restart_after_crash",
"value": "off",
"vartype": "bool"
},
{
"name": "synchronous_standby_names",
"value": "walproposer",
"vartype": "string"
},
{
"name": "shared_preload_libraries",
"value": "neon",
"vartype": "string"
},
{
"name": "neon.safekeepers",
"value": "safekeeper1:5454,safekeeper2:5454,safekeeper3:5454",
"vartype": "string"
},
{
"name": "neon.timeline_id",
"value": "TIMELINE_ID",
"vartype": "string"
},
{
"name": "neon.tenant_id",
"value": "TENANT_ID",
"vartype": "string"
},
{
"name": "neon.pageserver_connstring",
"value": "host=pageserver port=6400",
"vartype": "string"
},
{
"name": "max_replication_write_lag",
"value": "500MB",
"vartype": "string"
},
{
"name": "max_replication_flush_lag",
"value": "10GB",
"vartype": "string"
}
]
},
"delta_operations": [
]
}

View File

@@ -0,0 +1,200 @@
version: '3'
services:
etcd:
image: quay.io/coreos/etcd:v3.5.4
ports:
- 2379:2379
- 2380:2380
environment:
# This signifficantly speeds up etcd and we anyway don't data persistency there.
ETCD_UNSAFE_NO_FSYNC: "1"
command:
- "etcd"
- "--auto-compaction-mode=revision"
- "--auto-compaction-retention=1"
- "--name=etcd-cluster"
- "--initial-cluster-state=new"
- "--initial-cluster-token=etcd-cluster-1"
- "--initial-cluster=etcd-cluster=http://etcd:2380"
- "--initial-advertise-peer-urls=http://etcd:2380"
- "--advertise-client-urls=http://etcd:2379"
- "--listen-client-urls=http://0.0.0.0:2379"
- "--listen-peer-urls=http://0.0.0.0:2380"
- "--quota-backend-bytes=134217728" # 128 MB
minio:
image: quay.io/minio/minio:RELEASE.2022-10-20T00-55-09Z
ports:
- 9000:9000
- 9001:9001
environment:
- MINIO_ROOT_USER=minio
- MINIO_ROOT_PASSWORD=password
command: server /data --address :9000 --console-address ":9001"
minio_create_buckets:
image: minio/mc
environment:
- MINIO_ROOT_USER=minio
- MINIO_ROOT_PASSWORD=password
entrypoint:
- "/bin/sh"
- "-c"
command:
- "until (/usr/bin/mc alias set minio http://minio:9000 $$MINIO_ROOT_USER $$MINIO_ROOT_PASSWORD) do
echo 'Waiting to start minio...' && sleep 1;
done;
/usr/bin/mc mb minio/neon --region=eu-north-1;
exit 0;"
depends_on:
- minio
pageserver:
image: neondatabase/neon:${TAG:-latest}
environment:
- BROKER_ENDPOINT='http://etcd:2379'
- AWS_ACCESS_KEY_ID=minio
- AWS_SECRET_ACCESS_KEY=password
#- RUST_BACKTRACE=1
ports:
#- 6400:6400 # pg protocol handler
- 9898:9898 # http endpoints
entrypoint:
- "/bin/sh"
- "-c"
command:
- "/usr/local/bin/pageserver -D /data/.neon/
-c \"broker_endpoints=[$$BROKER_ENDPOINT]\"
-c \"listen_pg_addr='0.0.0.0:6400'\"
-c \"listen_http_addr='0.0.0.0:9898'\"
-c \"remote_storage={endpoint='http://minio:9000',
bucket_name='neon',
bucket_region='eu-north-1',
prefix_in_bucket='/pageserver/'}\""
depends_on:
- etcd
- minio_create_buckets
safekeeper1:
image: neondatabase/neon:${TAG:-latest}
environment:
- SAFEKEEPER_ADVERTISE_URL=safekeeper1:5454
- SAFEKEEPER_ID=1
- BROKER_ENDPOINT=http://etcd:2379
- AWS_ACCESS_KEY_ID=minio
- AWS_SECRET_ACCESS_KEY=password
#- RUST_BACKTRACE=1
ports:
#- 5454:5454 # pg protocol handler
- 7676:7676 # http endpoints
entrypoint:
- "/bin/sh"
- "-c"
command:
- "safekeeper --listen-pg=$$SAFEKEEPER_ADVERTISE_URL
--listen-http='0.0.0.0:7676'
--id=$$SAFEKEEPER_ID
--broker-endpoints=$$BROKER_ENDPOINT
-D /data
--remote-storage=\"{endpoint='http://minio:9000',
bucket_name='neon',
bucket_region='eu-north-1',
prefix_in_bucket='/safekeeper/'}\""
depends_on:
- etcd
- minio_create_buckets
safekeeper2:
image: neondatabase/neon:${TAG:-latest}
environment:
- SAFEKEEPER_ADVERTISE_URL=safekeeper2:5454
- SAFEKEEPER_ID=2
- BROKER_ENDPOINT=http://etcd:2379
- AWS_ACCESS_KEY_ID=minio
- AWS_SECRET_ACCESS_KEY=password
#- RUST_BACKTRACE=1
ports:
#- 5454:5454 # pg protocol handler
- 7677:7676 # http endpoints
entrypoint:
- "/bin/sh"
- "-c"
command:
- "safekeeper --listen-pg=$$SAFEKEEPER_ADVERTISE_URL
--listen-http='0.0.0.0:7676'
--id=$$SAFEKEEPER_ID
--broker-endpoints=$$BROKER_ENDPOINT
-D /data
--remote-storage=\"{endpoint='http://minio:9000',
bucket_name='neon',
bucket_region='eu-north-1',
prefix_in_bucket='/safekeeper/'}\""
depends_on:
- etcd
- minio_create_buckets
safekeeper3:
image: neondatabase/neon:${TAG:-latest}
environment:
- SAFEKEEPER_ADVERTISE_URL=safekeeper3:5454
- SAFEKEEPER_ID=3
- BROKER_ENDPOINT=http://etcd:2379
- AWS_ACCESS_KEY_ID=minio
- AWS_SECRET_ACCESS_KEY=password
#- RUST_BACKTRACE=1
ports:
#- 5454:5454 # pg protocol handler
- 7678:7676 # http endpoints
entrypoint:
- "/bin/sh"
- "-c"
command:
- "safekeeper --listen-pg=$$SAFEKEEPER_ADVERTISE_URL
--listen-http='0.0.0.0:7676'
--id=$$SAFEKEEPER_ID
--broker-endpoints=$$BROKER_ENDPOINT
-D /data
--remote-storage=\"{endpoint='http://minio:9000',
bucket_name='neon',
bucket_region='eu-north-1',
prefix_in_bucket='/safekeeper/'}\""
depends_on:
- etcd
- minio_create_buckets
compute:
build:
context: ./image/compute
args:
- COMPUTE_IMAGE=compute-node-v${PG_VERSION:-14}:${TAG:-latest}
- http_proxy=$http_proxy
- https_proxy=$https_proxy
environment:
- PG_VERSION=${PG_VERSION:-14}
#- RUST_BACKTRACE=1
volumes:
- ./compute/var/db/postgres/specs/:/var/db/postgres/specs/
- ./compute/shell/:/shell/
ports:
- 55433:55433 # pg protocol handler
- 3080:3080 # http endpoints
entrypoint:
- "/shell/compute.sh"
depends_on:
- safekeeper1
- safekeeper2
- safekeeper3
- pageserver
compute_is_ready:
image: postgres:latest
entrypoint:
- "/bin/bash"
- "-c"
command:
- "until pg_isready -h compute -p 55433 ; do
echo 'Waiting to start compute...' && sleep 1;
done"
depends_on:
- compute

View File

@@ -0,0 +1,10 @@
ARG COMPUTE_IMAGE=compute-node-v14:latest
FROM neondatabase/${COMPUTE_IMAGE}
USER root
RUN apt-get update && \
apt-get install -y curl \
jq \
netcat
USER postgres

View File

@@ -80,4 +80,6 @@
- [015-storage-messaging](rfcs/015-storage-messaging.md)
- [016-connection-routing](rfcs/016-connection-routing.md)
- [017-timeline-data-management](rfcs/017-timeline-data-management.md)
- [018-storage-messaging-2](rfcs/018-storage-messaging-2.md)
- [019-tenant-timeline-lifecycles](rfcs/019-tenant-timeline-lifecycles.md)
- [cluster-size-limits](rfcs/cluster-size-limits.md)

View File

@@ -18,3 +18,67 @@ We build all images after a successful `release` tests run and push automaticall
1. `neondatabase/compute-tools` and `neondatabase/compute-node`
2. `neondatabase/neon`
## Docker Compose example
You can see a [docker compose](https://docs.docker.com/compose/) example to create a neon cluster in [/docker-compose/docker-compose.yml](/docker-compose/docker-compose.yml). It creates the following conatainers.
- etcd x 1
- pageserver x 1
- safekeeper x 3
- compute x 1
- MinIO x 1 # This is Amazon S3 compatible object storage
### How to use
1. create containers
You can specify version of neon cluster using following environment values.
- PG_VERSION: postgres version for compute (default is 14)
- TAG: the tag version of [docker image](https://registry.hub.docker.com/r/neondatabase/neon/tags) (default is latest), which is tagged in [CI test](/.github/workflows/build_and_test.yml)
```
$ cd docker-compose/docker-compose.yml
$ docker-compose down # remove the conainers if exists
$ PG_VERSION=15 TAG=2221 docker-compose up --build -d # You can specify the postgres and image version
Creating network "dockercompose_default" with the default driver
Creating dockercompose_etcd3_1 ...
(...omit...)
```
2. connect compute node
```
$ echo "localhost:55433:postgres:cloud_admin:cloud_admin" >> ~/.pgpass
$ psql -h localhost -p 55433 -U cloud_admin
postgres=# CREATE TABLE t(key int primary key, value text);
CREATE TABLE
postgres=# insert into t values(1,1);
INSERT 0 1
postgres=# select * from t;
key | value
-----+-------
1 | 1
(1 row)
```
3. If you want to see the log, you can use `docker-compose logs` command.
```
# check the container name you want to see
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
d6968a5ae912 dockercompose_compute "/shell/compute.sh" 5 minutes ago Up 5 minutes 0.0.0.0:3080->3080/tcp, 0.0.0.0:55433->55433/tcp dockercompose_compute_1
(...omit...)
$ docker logs -f dockercompose_compute_1
2022-10-21 06:15:48.757 GMT [56] LOG: connection authorized: user=cloud_admin database=postgres application_name=psql
2022-10-21 06:17:00.307 GMT [56] LOG: [NEON_SMGR] libpagestore: connected to 'host=pageserver port=6400'
(...omit...)
```
4. If you want to see durable data in MinIO which is s3 compatible storage
Access http://localhost:9001 and sign in.
- Username: `minio`
- Password: `password`
You can see durable pages and WAL data in `neon` bucket.

View File

@@ -0,0 +1,163 @@
# Storage messaging
Safekeepers need to communicate to each other to
* Trim WAL on safekeepers;
* Decide on which SK should push WAL to the S3;
* Decide on when to shut down SK<->pageserver connection;
* Understand state of each other to perform peer recovery;
Pageservers need to communicate to safekeepers to decide which SK should provide
WAL to the pageserver.
This is an iteration on [015-storage-messaging](https://github.com/neondatabase/neon/blob/main/docs/rfcs/015-storage-messaging.md) describing current situation,
potential performance issue and ways to address it.
## Background
What we have currently is very close to etcd variant described in
015-storage-messaging. Basically, we have single `SkTimelineInfo` message
periodically sent by all safekeepers to etcd for each timeline.
* Safekeepers subscribe to it to learn status of peers (currently they subscribe to
'everything', but they can and should fetch data only for timelines they hold).
* Pageserver subscribes to it (separate watch per timeline) to learn safekeepers
positions; based on that, it decides from which safekeepers to pull WAL.
Also, safekeepers use etcd elections API to make sure only single safekeeper
offloads WAL.
It works, and callmemaybe is gone. However, this has a performance
hazard. Currently deployed etcd can do about 6k puts per second (using its own
`benchmark` tool); on my 6 core laptop, while running on tmpfs, this gets to
35k. Making benchmark closer to our usage [etcd watch bench](https://github.com/arssher/etcd-client/blob/watch-bench/examples/watch_bench.rs),
I get ~10k received messages per second with various number of publisher-subscribers
(laptop, tmpfs). Diving this by 12 (3 sks generate msg, 1 ps + 3 sk consume them) we
get about 800 active timelines, if message is sent each second. Not extremely
low, but quite reachable.
A lot of idle watches seem to be ok though -- which is good, as pageserver
subscribes to all its timelines regardless of their activity.
Also, running etcd with fsyncs disabled is messy -- data dir must be wiped on
each restart or there is a risk of corruption errors.
The reason is etcd making much more than what we need; it is a fault tolerant
store with strong consistency, but I claim all we need here is just simplest pub
sub with best effort delivery, because
* We already have centralized source of truth for long running data, like which
tlis are on which nodes -- the console.
* Momentary data (safekeeper/pageserver progress) doesn't make sense to persist.
Instead of putting each change to broker, expecting it to reliably deliver it
is better to just have constant flow of data for active timelines: 1) they
serve as natural heartbeats -- if node can't send, we shouldn't pull WAL from
it 2) it is simpler -- no need to track delivery to/from the broker.
Moreover, latency here is important: the faster we obtain fresh data, the
faster we can switch to proper safekeeper after failure.
* As for WAL offloading leader election, it is trivial to achieve through these
heartbeats -- just take suitable node through deterministic rule (min node
id). Once network is stable, this is a converging process (well, except
complicated failure topology, but even then making it converge is not
hard). Such elections bear some risk of several offloaders running
concurrently for a short period of time, but that's harmless.
Generally, if one needs strong consistency, electing leader per se is not
enough; it must be accompanied with number (logical clock ts), checked at
every action to track causality. s3 doesn't provide CAS, so it can't
differentiate old/new leader, this must be solved differently.
We could use etcd CAS (its most powerful/useful primitive actually) to issue
these leader numbers (and e.g. prefix files in s3), but currently I don't see
need for that.
Obviously best effort pub sub is much more simpler and performant; the one proposed is
## gRPC broker
I took tonic and [prototyped](https://github.com/neondatabase/neon/blob/asher/neon-broker/broker/src/broker.rs) the replacement of functionality we currently use
with grpc streams and tokio mpsc channels. The implementation description is at the file header.
It is just 500 lines of code and core functionality is complete. 1-1 pub sub
gives about 120k received messages per second; having multiple subscribers in
different connecitons quickly scales to 1 million received messages per second.
I had concerns about many concurrent streams in singe connection, but 2^20
subscribers still work (though eat memory, with 10 publishers 20GB are consumed;
in this implementation each publisher holds full copy of all subscribers). There
is `bench.rs` nearby which I used for testing.
`SkTimelineInfo` is wired here, but another message can be added (e.g. if
pageservers want to communicate with each other) with templating.
### Fault tolerance
Since such broker is stateless, we can run it under k8s. Or add proxying to
other members, with best-effort this is simple.
### Security implications
Communication happens in a private network that is not exposed to users;
additionaly we can add auth to the broker.
## Alternative: get existing pub-sub
We could take some existing pub sub solution, e.g. RabbitMQ, Redis. But in this
case IMV simplicity of our own outweights external dependency costs (RabbitMQ is
much more complicated and needs VM; Redis Rust client maintenance is not
ideal...). Also note that projects like CockroachDB and TiDB are based on gRPC
as well.
## Alternative: direct communication
Apart from being transport, broker solves one more task: discovery, i.e. letting
safekeepers and pageservers find each other. We can let safekeepers know, for
each timeline, both other safekeepers for this timeline and pageservers serving
it. In this case direct communication is possible:
- each safekeeper pushes to each other safekeeper status of timelines residing
on both of them, letting remove WAL, decide who offloads, decide on peer
recovery;
- each safekeeper pushes to each pageserver status of timelines residing on
both of them, letting pageserver choose from which sk to pull WAL;
It was mostly described in [014-safekeeper-gossip](https://github.com/neondatabase/neon/blob/main/docs/rfcs/014-safekeepers-gossip.md), but I want to recap on that.
The main pro is less one dependency: less moving parts, easier to run Neon
locally/manually, less places to monitor. Fault tolerance for broker disappears,
no kuber or something. To me this is a big thing.
Also (though not a big thing) idle watches for inactive timelines disappear:
naturally safekeepers learn about compute connection first and start pushing
status to pageserver(s), notifying it should pull.
Importantly, I think that eventually knowing and persisting peers and
pageservers on safekeepers is inevitable:
- Knowing peer safekeepers for the timeline is required for correct
automatic membership change -- new member set must be hardened on old
majority before proceeding. It is required to get rid of sync-safekeepers
as well (peer recovery up to flush_lsn).
- Knowing pageservers where the timeline is attached is needed to
1. Understand when to shut down activity on the timeline, i.e. push data to
the broker. We can have a lot of timelines sleeping quietly which
shouldn't occupy resources.
2. Preserve WAL for these (currently we offload to s3 and take it from there,
but serving locally is better, and we get one less condition on which WAL
can be removed from s3).
I suppose this membership data should be passed to safekeepers directly from the
console because
1. Console is the original source of this data, conceptually this is the
simplest way (rather than passing it through compute or something).
2. We already have similar code for deleting timeline on safekeepers
(and attaching/detaching timeline on pageserver), this is a typical
action -- queue operation against storage node and execute it until it
completes (or timeline is dropped).
Cons of direct communication are
- It is more complicated: each safekeeper should maintain set of peers it talks
to, and set of timelines for each such peer -- they ought to be multiplexed
into single connection.
- Totally, we have O(n^2) connections instead of O(n) with broker schema
(still O(n) on each node). However, these are relatively stable, async and
thus not very expensive, I don't think this is a big problem. Up to 10k
storage nodes I doubt connection overhead would be noticeable.
I'd use gRPC for direct communication, and in this sense gRPC based broker is a
step towards it.

View File

@@ -0,0 +1,91 @@
# Managing Tenant and Timeline lifecycles
## Summary
The pageserver has a Tenant object in memory for each tenant it manages, and a
Timeline for each timeline. There are a lot of tasks that operate on the tenants
and timelines with references to those objects. We have some mechanisms to track
which tasks are operating on each Tenant and Timeline, and to request them to
shutdown when a tenant or timeline is deleted, but it does not cover all uses,
and as a result we have many race conditions around tenant/timeline shutdown.
## Motivation
We have a bunch of race conditions that can produce weird errors and can be hard
to track down.
## Non Goals
This RFC only covers the problem of ensuring that a task/thread isn't operating
on a Tenant or Timeline. It does not cover what states, aside from Active and
non-Active, each Tenant and Timeline should have, or when exactly the transitions
should happen.
## Impacted components (e.g. pageserver, safekeeper, console, etc)
Pageserver. Although I wonder if the safekeeper should have a similar mechanism.
## Current situation
Most pageserver tasks of are managed by task_mgr.rs:
- LibpqEndpointListener
- HttpEndPointListener
- WalReceiverManager and -Connection
- GarbageCollector and Compaction
- InitialLogicalSizeCalculation
In addition to those tasks, the walreceiver performs some direct tokio::spawn
calls to spawn tasks that are not registered with 'task_mgr'. And all of these
tasks can spawn extra operations with tokio spawn_blocking.
Whenever a tenant or timeline is removed from the system, by pageserver
shutdown, delete_timeline or tenant-detach operation, we rely on the task
registry in 'task_mgr.rs' to wait until there are no tasks operating on the
tenant or timeline, before its Tenant/Timeline object is removed. That relies on
each task to register itself with the tenant/timeline ID in
'task_mgr.rs'. However, there are many gaps in that. For example,
GarbageCollection and Compaction tasks are registered with the tenant, but when
they proceed to operate on a particular timeline of the tenant, they don't
register with timeline ID. Because of that, the timeline can be deleted while GC
or compaction is running on it, causing failures in the GC or compaction (see
https://github.com/neondatabase/neon/issues/2442).
Another problem is that the task registry only works for tokio Tasks. There is
no way to register a piece of code that runs inside spawn_blocking(), for
example.
## Proposed implementation
This "voluntary" registration of tasks is fragile. Let's use Rust language features
to enforce that a tenant/timeline cannot be removed from the system when there is
still some code operating on it.
Let's introduce new Guard objects for Tenant and Timeline, and do all actions through
the Guard object. Something like:
TenantActiveGuard: Guard object over Arc<Tenant>. When you acquire the guard,
the code checks that the tenant is in Active state. If it's not, you get an
error. You can change the state of the tenant to Stopping while there are
ActiveTenantGuard objects still on it, to prevent new ActiveTenantGuards from
being acquired, but the Tenant cannot be removed until all the guards are gone.
TenantMaintenanceGuard: Like ActiveTenantGuard, but can be held even when the
tenant is not in Active state. Used for operations like attach/detach. Perhaps
allow only one such guard on a Tenant at a time.
Similarly for Timelines. We don't currentl have a "state" on Timeline, but I think
we need at least two states: Active and Stopping. The Stopping state is used at
deletion, to prevent new TimelineActiveGuards from appearing, while you wait for
existing TimelineActiveGuards to die out.
The shutdown-signaling, using shutdown_watcher() and is_shutdown_requested(),
probably also needs changes to deal with the new Guards. The rule is that if you
have a TenantActiveGuard, and the tenant's state changes from Active to
Stopping, the is_shutdown_requested() function should return true, and
shutdown_watcher() future should return.
This signaling doesn't neessarily need to cover all cases. For example, if you
have a block of code in spawn_blocking(), it might be acceptable if
is_shutdown_requested() doesn't return true even though the tenant is in
Stopping state, as long as the code finishes reasonably fast.

View File

@@ -96,7 +96,7 @@ A single virtual environment with all dependencies is described in the single `P
sudo apt install python3.9
```
- Install `poetry`
- Exact version of `poetry` is not important, see installation instructions available at poetry's [website](https://python-poetry.org/docs/#installation)`.
- Exact version of `poetry` is not important, see installation instructions available at poetry's [website](https://python-poetry.org/docs/#installation).
- Install dependencies via `./scripts/pysync`.
- Note that CI uses specific Python version (look for `PYTHON_VERSION` [here](https://github.com/neondatabase/docker-images/blob/main/rust/Dockerfile))
so if you have different version some linting tools can yield different result locally vs in the CI.

View File

@@ -1,111 +0,0 @@
#!/usr/bin/env python3
import argparse
import json
import os
import subprocess
import sys
import textwrap
import uuid
from pathlib import Path
import testgres
def run_command(args):
print('> Cmd:', ' '.join(args))
p = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, text=True)
ret = p.wait()
output = p.stdout.read().strip()
if output:
print(textwrap.indent(output, '>> '))
if ret != 0:
raise subprocess.CalledProcessError(ret, args)
def make_tarfile(output_filename, source_dir):
print("* Packing the backup into a tarball")
cmd = ["tar", r"--transform=s/\.\///", "-C", str(source_dir), "-cf", str(output_filename), "."]
run_command(cmd)
def create_tenant(tenant_id):
print("* Creating a new tenant")
cmd = ["neon_local", "tenant", "create", f"--tenant-id={tenant_id}"]
run_command(cmd)
def import_backup(args, backup_dir: Path):
tar = Path('/tmp/base.tar')
make_tarfile(tar, backup_dir / 'data')
print("* Importing the timeline into the pageserver")
manifest = json.loads((backup_dir / "data" / "backup_manifest").read_text())
start_lsn = manifest["WAL-Ranges"][0]["Start-LSN"]
end_lsn = manifest["WAL-Ranges"][0]["End-LSN"]
print("> LSNs:", start_lsn, end_lsn)
cmd = (
"neon_local timeline import "
f"--tenant-id {args.tenant_id} "
f"--base-lsn {start_lsn} "
f"--end-lsn {end_lsn} "
f"--base-tarfile {tar} "
f"--timeline-id {args.timeline_id} "
f"--node-name {args.node}"
)
run_command(cmd.split())
def debug_prints(node):
tuples = node.execute("table foo")
oid = node.execute("select 'foo'::regclass::oid")[0][0]
print("> foo's tuples:", tuples, "&", "oid:", oid)
print("> DBs:", node.execute("select oid, datname from pg_database"))
def main(args):
print("* Creating a node")
node = testgres.get_new_node()
node.init(unix_sockets=False, allow_streaming=True).start()
node.execute("create table foo as select 1")
debug_prints(node)
# node.pgbench_init(scale=1)
print("* Creating a backup")
backup = node.backup()
backup_dir = Path(backup.base_dir)
print("> Backup dir:", backup_dir)
# pr = backup.spawn_primary().start()
# debug_prints(pr)
# exit(1)
create_tenant(args.tenant_id)
import_backup(args, backup_dir)
print("> Tenant:", args.tenant_id)
print("> Timeline:", args.timeline_id)
print("> Node:", args.node)
print("* Starting postgres")
cmd = ["neon_local", "pg", "start", f"--tenant-id={args.tenant_id}", f"--timeline-id={args.timeline_id}", args.node]
run_command(cmd)
print("* Opening psql session...")
cmd = ["psql", f"host=127.0.0.1 port=55432 user={os.getlogin()} dbname=postgres"]
subprocess.call(cmd)
if __name__ == "__main__":
tenant_id = uuid.uuid4().hex
parser = argparse.ArgumentParser()
parser.add_argument("--tenant-id", default=tenant_id)
parser.add_argument("--timeline-id", default=tenant_id)
parser.add_argument("node")
args = parser.parse_args(sys.argv[1:])
main(args)

View File

@@ -8,7 +8,7 @@
regex = "1.4.5"
serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
serde_with = "1.12.0"
serde_with = "2.0"
once_cell = "1.13.0"
utils = { path = "../utils" }

View File

@@ -29,6 +29,9 @@ pub struct SkTimelineInfo {
#[serde_as(as = "Option<DisplayFromStr>")]
#[serde(default)]
pub peer_horizon_lsn: Option<Lsn>,
#[serde_as(as = "Option<DisplayFromStr>")]
#[serde(default)]
pub local_start_lsn: Option<Lsn>,
/// A connection string to use for WAL receiving.
#[serde(default)]
pub safekeeper_connstr: Option<String>,

View File

@@ -3,7 +3,7 @@
//! Otherwise, we might not see all metrics registered via
//! a default registry.
use once_cell::sync::Lazy;
use prometheus::core::{AtomicU64, GenericGauge, GenericGaugeVec};
use prometheus::core::{AtomicU64, Collector, GenericGauge, GenericGaugeVec};
pub use prometheus::opts;
pub use prometheus::register;
pub use prometheus::{core, default_registry, proto};
@@ -17,6 +17,7 @@ pub use prometheus::{register_int_counter_vec, IntCounterVec};
pub use prometheus::{register_int_gauge, IntGauge};
pub use prometheus::{register_int_gauge_vec, IntGaugeVec};
pub use prometheus::{Encoder, TextEncoder};
use prometheus::{Registry, Result};
mod wrappers;
pub use wrappers::{CountedReader, CountedWriter};
@@ -32,13 +33,27 @@ macro_rules! register_uint_gauge_vec {
}};
}
/// Special internal registry, to collect metrics independently from the default registry.
/// Was introduced to fix deadlock with lazy registration of metrics in the default registry.
static INTERNAL_REGISTRY: Lazy<Registry> = Lazy::new(Registry::new);
/// Register a collector in the internal registry. MUST be called before the first call to `gather()`.
/// Otherwise, we can have a deadlock in the `gather()` call, trying to register a new collector
/// while holding the lock.
pub fn register_internal(c: Box<dyn Collector>) -> Result<()> {
INTERNAL_REGISTRY.register(c)
}
/// Gathers all Prometheus metrics and records the I/O stats just before that.
///
/// Metrics gathering is a relatively simple and standalone operation, so
/// it might be fine to do it this way to keep things simple.
pub fn gather() -> Vec<prometheus::proto::MetricFamily> {
update_rusage_metrics();
prometheus::gather()
let mut mfs = prometheus::gather();
let mut internal_mfs = INTERNAL_REGISTRY.gather();
mfs.append(&mut internal_mfs);
mfs
}
static DISK_IO_BYTES: Lazy<IntGaugeVec> = Lazy::new(|| {
@@ -62,6 +77,16 @@ pub const DISK_WRITE_SECONDS_BUCKETS: &[f64] = &[
0.000_050, 0.000_100, 0.000_500, 0.001, 0.003, 0.005, 0.01, 0.05, 0.1, 0.3, 0.5,
];
pub fn set_build_info_metric(revision: &str) {
let metric = register_int_gauge_vec!(
"libmetrics_build_info",
"Build/version information",
&["revision"]
)
.expect("Failed to register build info metric");
metric.with_label_values(&[revision]).set(1);
}
// Records I/O stats in a "cross-platform" way.
// Compiles both on macOS and Linux, but current macOS implementation always returns 0 as values for I/O stats.
// An alternative is to read procfs (`/proc/[pid]/io`) which does not work under macOS at all, hence abandoned.

View File

@@ -0,0 +1,15 @@
[package]
name = "pageserver_api"
version = "0.1.0"
edition = "2021"
[dependencies]
serde = { version = "1.0", features = ["derive"] }
serde_with = "2.0"
const_format = "0.2.21"
anyhow = { version = "1.0", features = ["backtrace"] }
bytes = "1.0.1"
utils = { path = "../utils" }
postgres_ffi = { path = "../postgres_ffi" }
workspace_hack = { version = "0.1", path = "../../workspace_hack" }

View File

@@ -0,0 +1,10 @@
use const_format::formatcp;
/// Public API types
pub mod models;
pub mod reltag;
pub const DEFAULT_PG_LISTEN_PORT: u16 = 64000;
pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 9898;
pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");

View File

@@ -0,0 +1,382 @@
use std::num::NonZeroU64;
use serde::{Deserialize, Serialize};
use serde_with::{serde_as, DisplayFromStr};
use utils::{
id::{NodeId, TenantId, TimelineId},
lsn::Lsn,
};
use crate::reltag::RelTag;
use anyhow::bail;
use bytes::{Buf, BufMut, Bytes, BytesMut};
/// A state of a tenant in pageserver's memory.
#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub enum TenantState {
/// Tenant is fully operational, its background jobs might be running or not.
Active { background_jobs_running: bool },
/// A tenant is recognized by pageserver, but not yet ready to operate:
/// e.g. not present locally and being downloaded or being read into memory from the file system.
Paused,
/// A tenant is recognized by the pageserver, but no longer used for any operations, as failed to get activated.
Broken,
}
/// A state of a timeline in pageserver's memory.
#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub enum TimelineState {
/// Timeline is fully operational, its background jobs are running.
Active,
/// A timeline is recognized by pageserver, but not yet ready to operate.
/// The status indicates, that the timeline could eventually go back to Active automatically:
/// for example, if the owning tenant goes back to Active again.
Suspended,
/// A timeline is recognized by pageserver, but not yet ready to operate and not allowed to
/// automatically become Active after certain events: only a management call can change this status.
Paused,
/// A timeline is recognized by the pageserver, but no longer used for any operations, as failed to get activated.
Broken,
}
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TimelineCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_start_lsn: Option<Lsn>,
pub pg_version: Option<u32>,
}
#[serde_as]
#[derive(Serialize, Deserialize, Default)]
pub struct TenantCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_tenant_id: Option<TenantId>,
pub checkpoint_distance: Option<u64>,
pub checkpoint_timeout: Option<String>,
pub compaction_target_size: Option<u64>,
pub compaction_period: Option<String>,
pub compaction_threshold: Option<usize>,
pub gc_horizon: Option<u64>,
pub gc_period: Option<String>,
pub image_creation_threshold: Option<usize>,
pub pitr_interval: Option<String>,
pub walreceiver_connect_timeout: Option<String>,
pub lagging_wal_timeout: Option<String>,
pub max_lsn_wal_lag: Option<NonZeroU64>,
}
#[serde_as]
#[derive(Serialize, Deserialize)]
#[serde(transparent)]
pub struct TenantCreateResponse(#[serde_as(as = "DisplayFromStr")] pub TenantId);
#[derive(Serialize)]
pub struct StatusResponse {
pub id: NodeId,
}
impl TenantCreateRequest {
pub fn new(new_tenant_id: Option<TenantId>) -> TenantCreateRequest {
TenantCreateRequest {
new_tenant_id,
..Default::default()
}
}
}
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TenantConfigRequest {
pub tenant_id: TenantId,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub checkpoint_distance: Option<u64>,
pub checkpoint_timeout: Option<String>,
pub compaction_target_size: Option<u64>,
pub compaction_period: Option<String>,
pub compaction_threshold: Option<usize>,
pub gc_horizon: Option<u64>,
pub gc_period: Option<String>,
pub image_creation_threshold: Option<usize>,
pub pitr_interval: Option<String>,
pub walreceiver_connect_timeout: Option<String>,
pub lagging_wal_timeout: Option<String>,
pub max_lsn_wal_lag: Option<NonZeroU64>,
}
impl TenantConfigRequest {
pub fn new(tenant_id: TenantId) -> TenantConfigRequest {
TenantConfigRequest {
tenant_id,
checkpoint_distance: None,
checkpoint_timeout: None,
compaction_target_size: None,
compaction_period: None,
compaction_threshold: None,
gc_horizon: None,
gc_period: None,
image_creation_threshold: None,
pitr_interval: None,
walreceiver_connect_timeout: None,
lagging_wal_timeout: None,
max_lsn_wal_lag: None,
}
}
}
#[serde_as]
#[derive(Serialize, Deserialize, Clone)]
pub struct TenantInfo {
#[serde_as(as = "DisplayFromStr")]
pub id: TenantId,
pub state: TenantState,
pub current_physical_size: Option<u64>, // physical size is only included in `tenant_status` endpoint
pub has_in_progress_downloads: Option<bool>,
}
/// This represents the output of the "timeline_detail" and "timeline_list" API calls.
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct TimelineInfo {
#[serde_as(as = "DisplayFromStr")]
pub tenant_id: TenantId,
#[serde_as(as = "DisplayFromStr")]
pub timeline_id: TimelineId,
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<TimelineId>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_lsn: Option<Lsn>,
#[serde_as(as = "DisplayFromStr")]
pub last_record_lsn: Lsn,
#[serde_as(as = "Option<DisplayFromStr>")]
pub prev_record_lsn: Option<Lsn>,
#[serde_as(as = "DisplayFromStr")]
pub latest_gc_cutoff_lsn: Lsn,
#[serde_as(as = "DisplayFromStr")]
pub disk_consistent_lsn: Lsn,
pub current_logical_size: Option<u64>, // is None when timeline is Unloaded
pub current_physical_size: Option<u64>, // is None when timeline is Unloaded
pub current_logical_size_non_incremental: Option<u64>,
pub current_physical_size_non_incremental: Option<u64>,
pub wal_source_connstr: Option<String>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub last_received_msg_lsn: Option<Lsn>,
/// the timestamp (in microseconds) of the last received message
pub last_received_msg_ts: Option<u128>,
pub pg_version: u32,
#[serde_as(as = "Option<DisplayFromStr>")]
pub remote_consistent_lsn: Option<Lsn>,
pub awaits_download: bool,
pub state: TimelineState,
// Some of the above fields are duplicated in 'local' and 'remote', for backwards-
// compatility with older clients.
pub local: LocalTimelineInfo,
pub remote: RemoteTimelineInfo,
}
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct LocalTimelineInfo {
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<TimelineId>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_lsn: Option<Lsn>,
pub current_logical_size: Option<u64>, // is None when timeline is Unloaded
pub current_physical_size: Option<u64>, // is None when timeline is Unloaded
}
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct RemoteTimelineInfo {
#[serde_as(as = "Option<DisplayFromStr>")]
pub remote_consistent_lsn: Option<Lsn>,
}
pub type ConfigureFailpointsRequest = Vec<FailpointConfig>;
/// Information for configuring a single fail point
#[derive(Debug, Serialize, Deserialize)]
pub struct FailpointConfig {
/// Name of the fail point
pub name: String,
/// List of actions to take, using the format described in `fail::cfg`
///
/// We also support `actions = "exit"` to cause the fail point to immediately exit.
pub actions: String,
}
#[derive(Debug, Serialize, Deserialize)]
pub struct TimelineGcRequest {
pub gc_horizon: Option<u64>,
}
// Wrapped in libpq CopyData
pub enum PagestreamFeMessage {
Exists(PagestreamExistsRequest),
Nblocks(PagestreamNblocksRequest),
GetPage(PagestreamGetPageRequest),
DbSize(PagestreamDbSizeRequest),
}
// Wrapped in libpq CopyData
pub enum PagestreamBeMessage {
Exists(PagestreamExistsResponse),
Nblocks(PagestreamNblocksResponse),
GetPage(PagestreamGetPageResponse),
Error(PagestreamErrorResponse),
DbSize(PagestreamDbSizeResponse),
}
#[derive(Debug)]
pub struct PagestreamExistsRequest {
pub latest: bool,
pub lsn: Lsn,
pub rel: RelTag,
}
#[derive(Debug)]
pub struct PagestreamNblocksRequest {
pub latest: bool,
pub lsn: Lsn,
pub rel: RelTag,
}
#[derive(Debug)]
pub struct PagestreamGetPageRequest {
pub latest: bool,
pub lsn: Lsn,
pub rel: RelTag,
pub blkno: u32,
}
#[derive(Debug)]
pub struct PagestreamDbSizeRequest {
pub latest: bool,
pub lsn: Lsn,
pub dbnode: u32,
}
#[derive(Debug)]
pub struct PagestreamExistsResponse {
pub exists: bool,
}
#[derive(Debug)]
pub struct PagestreamNblocksResponse {
pub n_blocks: u32,
}
#[derive(Debug)]
pub struct PagestreamGetPageResponse {
pub page: Bytes,
}
#[derive(Debug)]
pub struct PagestreamErrorResponse {
pub message: String,
}
#[derive(Debug)]
pub struct PagestreamDbSizeResponse {
pub db_size: i64,
}
impl PagestreamFeMessage {
pub fn parse(mut body: Bytes) -> anyhow::Result<PagestreamFeMessage> {
// TODO these gets can fail
// these correspond to the NeonMessageTag enum in pagestore_client.h
//
// TODO: consider using protobuf or serde bincode for less error prone
// serialization.
let msg_tag = body.get_u8();
match msg_tag {
0 => Ok(PagestreamFeMessage::Exists(PagestreamExistsRequest {
latest: body.get_u8() != 0,
lsn: Lsn::from(body.get_u64()),
rel: RelTag {
spcnode: body.get_u32(),
dbnode: body.get_u32(),
relnode: body.get_u32(),
forknum: body.get_u8(),
},
})),
1 => Ok(PagestreamFeMessage::Nblocks(PagestreamNblocksRequest {
latest: body.get_u8() != 0,
lsn: Lsn::from(body.get_u64()),
rel: RelTag {
spcnode: body.get_u32(),
dbnode: body.get_u32(),
relnode: body.get_u32(),
forknum: body.get_u8(),
},
})),
2 => Ok(PagestreamFeMessage::GetPage(PagestreamGetPageRequest {
latest: body.get_u8() != 0,
lsn: Lsn::from(body.get_u64()),
rel: RelTag {
spcnode: body.get_u32(),
dbnode: body.get_u32(),
relnode: body.get_u32(),
forknum: body.get_u8(),
},
blkno: body.get_u32(),
})),
3 => Ok(PagestreamFeMessage::DbSize(PagestreamDbSizeRequest {
latest: body.get_u8() != 0,
lsn: Lsn::from(body.get_u64()),
dbnode: body.get_u32(),
})),
_ => bail!("unknown smgr message tag: {},'{:?}'", msg_tag, body),
}
}
}
impl PagestreamBeMessage {
pub fn serialize(&self) -> Bytes {
let mut bytes = BytesMut::new();
match self {
Self::Exists(resp) => {
bytes.put_u8(100); /* tag from pagestore_client.h */
bytes.put_u8(resp.exists as u8);
}
Self::Nblocks(resp) => {
bytes.put_u8(101); /* tag from pagestore_client.h */
bytes.put_u32(resp.n_blocks);
}
Self::GetPage(resp) => {
bytes.put_u8(102); /* tag from pagestore_client.h */
bytes.put(&resp.page[..]);
}
Self::Error(resp) => {
bytes.put_u8(103); /* tag from pagestore_client.h */
bytes.put(resp.message.as_bytes());
bytes.put_u8(0); // null terminator
}
Self::DbSize(resp) => {
bytes.put_u8(104); /* tag from pagestore_client.h */
bytes.put_i64(resp.db_size);
}
}
bytes.into()
}
}

View File

@@ -13,7 +13,7 @@ crc32c = "0.6.0"
hex = "0.4.3"
once_cell = "1.13.0"
log = "0.4.14"
memoffset = "0.6.2"
memoffset = "0.7"
thiserror = "1.0"
serde = { version = "1.0", features = ["derive"] }
utils = { path = "../utils" }
@@ -26,4 +26,4 @@ wal_craft = { path = "wal_craft" }
[build-dependencies]
anyhow = "1.0"
bindgen = "0.60.1"
bindgen = "0.61"

View File

@@ -3,9 +3,11 @@
#![allow(non_snake_case)]
// bindgen creates some unsafe code with no doc comments.
#![allow(clippy::missing_safety_doc)]
// suppress warnings on rust 1.53 due to bindgen unit tests.
// https://github.com/rust-lang/rust-bindgen/issues/1651
#![allow(deref_nullptr)]
// noted at 1.63 that in many cases there's a u32 -> u32 transmutes in bindgen code.
#![allow(clippy::useless_transmute)]
// modules included with the postgres_ffi macro depend on the types of the specific version's
// types, and trigger a too eager lint.
#![allow(clippy::duplicate_mod)]
use bytes::Bytes;
use utils::bin_ser::SerializeError;

View File

@@ -57,12 +57,10 @@ pub const SIZE_OF_XLOG_RECORD_DATA_HEADER_SHORT: usize = 1 * 2;
/// in order to let CLOG_TRUNCATE mechanism correctly extend CLOG.
const XID_CHECKPOINT_INTERVAL: u32 = 1024;
#[allow(non_snake_case)]
pub fn XLogSegmentsPerXLogId(wal_segsz_bytes: usize) -> XLogSegNo {
(0x100000000u64 / wal_segsz_bytes as u64) as XLogSegNo
}
#[allow(non_snake_case)]
pub fn XLogSegNoOffsetToRecPtr(
segno: XLogSegNo,
offset: u32,
@@ -71,7 +69,6 @@ pub fn XLogSegNoOffsetToRecPtr(
segno * (wal_segsz_bytes as u64) + (offset as u64)
}
#[allow(non_snake_case)]
pub fn XLogFileName(tli: TimeLineID, logSegNo: XLogSegNo, wal_segsz_bytes: usize) -> String {
format!(
"{:>08X}{:>08X}{:>08X}",
@@ -81,7 +78,6 @@ pub fn XLogFileName(tli: TimeLineID, logSegNo: XLogSegNo, wal_segsz_bytes: usize
)
}
#[allow(non_snake_case)]
pub fn XLogFromFileName(fname: &str, wal_seg_size: usize) -> (XLogSegNo, TimeLineID) {
let tli = u32::from_str_radix(&fname[0..8], 16).unwrap();
let log = u32::from_str_radix(&fname[8..16], 16).unwrap() as XLogSegNo;
@@ -89,12 +85,10 @@ pub fn XLogFromFileName(fname: &str, wal_seg_size: usize) -> (XLogSegNo, TimeLin
(log * XLogSegmentsPerXLogId(wal_seg_size) + seg, tli)
}
#[allow(non_snake_case)]
pub fn IsXLogFileName(fname: &str) -> bool {
return fname.len() == XLOG_FNAME_LEN && fname.chars().all(|c| c.is_ascii_hexdigit());
}
#[allow(non_snake_case)]
pub fn IsPartialXLogFileName(fname: &str) -> bool {
fname.ends_with(".partial") && IsXLogFileName(&fname[0..fname.len() - 8])
}

View File

@@ -7,7 +7,7 @@ edition = "2021"
[dependencies]
anyhow = "1.0"
clap = "3.0"
clap = "4.0"
env_logger = "0.9"
log = "0.4"
once_cell = "1.13.0"

View File

@@ -1,68 +1,19 @@
use anyhow::*;
use clap::{App, Arg, ArgMatches};
use std::str::FromStr;
use clap::{value_parser, Arg, ArgMatches, Command};
use std::{path::PathBuf, str::FromStr};
use wal_craft::*;
fn main() -> Result<()> {
env_logger::Builder::from_env(env_logger::Env::default().default_filter_or("wal_craft=info"))
.init();
let type_arg = &Arg::new("type")
.takes_value(true)
.help("Type of WAL to craft")
.possible_values([
Simple::NAME,
LastWalRecordXlogSwitch::NAME,
LastWalRecordXlogSwitchEndsOnPageBoundary::NAME,
WalRecordCrossingSegmentFollowedBySmallOne::NAME,
LastWalRecordCrossingSegment::NAME,
])
.required(true);
let arg_matches = App::new("Postgres WAL crafter")
.about("Crafts Postgres databases with specific WAL properties")
.subcommand(
App::new("print-postgres-config")
.about("Print the configuration required for PostgreSQL server before running this script")
)
.subcommand(
App::new("with-initdb")
.about("Craft WAL in a new data directory first initialized with initdb")
.arg(type_arg)
.arg(
Arg::new("datadir")
.takes_value(true)
.help("Data directory for the Postgres server")
.required(true)
)
.arg(
Arg::new("pg-distrib-dir")
.long("pg-distrib-dir")
.takes_value(true)
.help("Directory with Postgres distributions (bin and lib directories, e.g. pg_install containing subpath `v14/bin/postgresql`)")
.default_value("/usr/local")
)
.arg(
Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.required(true)
.takes_value(true)
)
)
.subcommand(
App::new("in-existing")
.about("Craft WAL at an existing recently created Postgres database. Note that server may append new WAL entries on shutdown.")
.arg(type_arg)
.arg(
Arg::new("connection")
.takes_value(true)
.help("Connection string to the Postgres database to populate")
.required(true)
)
)
.get_matches();
let arg_matches = cli().get_matches();
let wal_craft = |arg_matches: &ArgMatches, client| {
let (intermediate_lsns, end_of_wal_lsn) = match arg_matches.value_of("type").unwrap() {
let (intermediate_lsns, end_of_wal_lsn) = match arg_matches
.get_one::<String>("type")
.map(|s| s.as_str())
.context("'type' is required")?
{
Simple::NAME => Simple::craft(client)?,
LastWalRecordXlogSwitch::NAME => LastWalRecordXlogSwitch::craft(client)?,
LastWalRecordXlogSwitchEndsOnPageBoundary::NAME => {
@@ -72,12 +23,12 @@ fn main() -> Result<()> {
WalRecordCrossingSegmentFollowedBySmallOne::craft(client)?
}
LastWalRecordCrossingSegment::NAME => LastWalRecordCrossingSegment::craft(client)?,
a => panic!("Unknown --type argument: {}", a),
a => panic!("Unknown --type argument: {a}"),
};
for lsn in intermediate_lsns {
println!("intermediate_lsn = {}", lsn);
println!("intermediate_lsn = {lsn}");
}
println!("end_of_wal = {}", end_of_wal_lsn);
println!("end_of_wal = {end_of_wal_lsn}");
Ok(())
};
@@ -85,20 +36,24 @@ fn main() -> Result<()> {
None => panic!("No subcommand provided"),
Some(("print-postgres-config", _)) => {
for cfg in REQUIRED_POSTGRES_CONFIG.iter() {
println!("{}", cfg);
println!("{cfg}");
}
Ok(())
}
Some(("with-initdb", arg_matches)) => {
let cfg = Conf {
pg_version: arg_matches
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?,
pg_distrib_dir: arg_matches.value_of("pg-distrib-dir").unwrap().into(),
datadir: arg_matches.value_of("datadir").unwrap().into(),
pg_version: *arg_matches
.get_one::<u32>("pg-version")
.context("'pg-version' is required")?,
pg_distrib_dir: arg_matches
.get_one::<PathBuf>("pg-distrib-dir")
.context("'pg-distrib-dir' is required")?
.to_owned(),
datadir: arg_matches
.get_one::<PathBuf>("datadir")
.context("'datadir' is required")?
.to_owned(),
};
cfg.initdb()?;
let srv = cfg.start_server()?;
@@ -108,9 +63,77 @@ fn main() -> Result<()> {
}
Some(("in-existing", arg_matches)) => wal_craft(
arg_matches,
&mut postgres::Config::from_str(arg_matches.value_of("connection").unwrap())?
.connect(postgres::NoTls)?,
&mut postgres::Config::from_str(
arg_matches
.get_one::<String>("connection")
.context("'connection' is required")?,
)
.context(
"'connection' argument value could not be parsed as a postgres connection string",
)?
.connect(postgres::NoTls)?,
),
Some(_) => panic!("Unknown subcommand"),
}
}
fn cli() -> Command {
let type_arg = &Arg::new("type")
.help("Type of WAL to craft")
.value_parser([
Simple::NAME,
LastWalRecordXlogSwitch::NAME,
LastWalRecordXlogSwitchEndsOnPageBoundary::NAME,
WalRecordCrossingSegmentFollowedBySmallOne::NAME,
LastWalRecordCrossingSegment::NAME,
])
.required(true);
Command::new("Postgres WAL crafter")
.about("Crafts Postgres databases with specific WAL properties")
.subcommand(
Command::new("print-postgres-config")
.about("Print the configuration required for PostgreSQL server before running this script")
)
.subcommand(
Command::new("with-initdb")
.about("Craft WAL in a new data directory first initialized with initdb")
.arg(type_arg)
.arg(
Arg::new("datadir")
.help("Data directory for the Postgres server")
.value_parser(value_parser!(PathBuf))
.required(true)
)
.arg(
Arg::new("pg-distrib-dir")
.long("pg-distrib-dir")
.value_parser(value_parser!(PathBuf))
.help("Directory with Postgres distributions (bin and lib directories, e.g. pg_install containing subpath `v14/bin/postgresql`)")
.default_value("/usr/local")
)
.arg(
Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.value_parser(value_parser!(u32))
.required(true)
)
)
.subcommand(
Command::new("in-existing")
.about("Craft WAL at an existing recently created Postgres database. Note that server may append new WAL entries on shutdown.")
.arg(type_arg)
.arg(
Arg::new("connection")
.help("Connection string to the Postgres database to populate")
.required(true)
)
)
}
#[test]
fn verify_cli() {
cli().debug_assert();
}

View File

@@ -37,22 +37,22 @@ pub static REQUIRED_POSTGRES_CONFIG: Lazy<Vec<&'static str>> = Lazy::new(|| {
});
impl Conf {
pub fn pg_distrib_dir(&self) -> PathBuf {
pub fn pg_distrib_dir(&self) -> anyhow::Result<PathBuf> {
let path = self.pg_distrib_dir.clone();
match self.pg_version {
14 => path.join(format!("v{}", self.pg_version)),
15 => path.join(format!("v{}", self.pg_version)),
_ => panic!("Unsupported postgres version: {}", self.pg_version),
14 => Ok(path.join(format!("v{}", self.pg_version))),
15 => Ok(path.join(format!("v{}", self.pg_version))),
_ => bail!("Unsupported postgres version: {}", self.pg_version),
}
}
fn pg_bin_dir(&self) -> PathBuf {
self.pg_distrib_dir().join("bin")
fn pg_bin_dir(&self) -> anyhow::Result<PathBuf> {
Ok(self.pg_distrib_dir()?.join("bin"))
}
fn pg_lib_dir(&self) -> PathBuf {
self.pg_distrib_dir().join("lib")
fn pg_lib_dir(&self) -> anyhow::Result<PathBuf> {
Ok(self.pg_distrib_dir()?.join("lib"))
}
pub fn wal_dir(&self) -> PathBuf {
@@ -60,12 +60,12 @@ impl Conf {
}
fn new_pg_command(&self, command: impl AsRef<Path>) -> Result<Command> {
let path = self.pg_bin_dir().join(command);
let path = self.pg_bin_dir()?.join(command);
ensure!(path.exists(), "Command {:?} does not exist", path);
let mut cmd = Command::new(path);
cmd.env_clear()
.env("LD_LIBRARY_PATH", self.pg_lib_dir())
.env("DYLD_LIBRARY_PATH", self.pg_lib_dir());
.env("LD_LIBRARY_PATH", self.pg_lib_dir()?)
.env("DYLD_LIBRARY_PATH", self.pg_lib_dir()?);
Ok(cmd)
}

View File

@@ -15,7 +15,7 @@ serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
tokio = { version = "1.17", features = ["sync", "macros", "fs", "io-util"] }
tokio-util = { version = "0.7", features = ["io"] }
toml_edit = { version = "0.13", features = ["easy"] }
toml_edit = { version = "0.14", features = ["easy"] }
tracing = "0.1.27"
workspace_hack = { version = "0.1", path = "../../workspace_hack" }

View File

@@ -16,7 +16,7 @@ use tokio::{
io::{self, AsyncReadExt, AsyncSeekExt, AsyncWriteExt},
};
use tracing::*;
use utils::crashsafe_dir::path_with_suffix_extension;
use utils::crashsafe::path_with_suffix_extension;
use crate::{Download, DownloadError, RemoteObjectId};

View File

@@ -0,0 +1,12 @@
[package]
name = "safekeeper_api"
version = "0.1.0"
edition = "2021"
[dependencies]
serde = { version = "1.0", features = ["derive"] }
serde_with = "2.0"
const_format = "0.2.21"
utils = { path = "../utils" }
workspace_hack = { version = "0.1", path = "../../workspace_hack" }

View File

@@ -0,0 +1,10 @@
use const_format::formatcp;
/// Public API types
pub mod models;
pub const DEFAULT_PG_LISTEN_PORT: u16 = 5454;
pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 7676;
pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");

View File

@@ -0,0 +1,24 @@
use serde::{Deserialize, Serialize};
use serde_with::{serde_as, DisplayFromStr};
use utils::{
id::{NodeId, TenantId, TimelineId},
lsn::Lsn,
};
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TimelineCreateRequest {
#[serde_as(as = "DisplayFromStr")]
pub tenant_id: TenantId,
#[serde_as(as = "DisplayFromStr")]
pub timeline_id: TimelineId,
pub peer_ids: Option<Vec<NodeId>>,
pub pg_version: u32,
pub system_id: Option<u64>,
pub wal_seg_size: Option<u32>,
#[serde_as(as = "DisplayFromStr")]
pub commit_lsn: Lsn,
// If not passed, it is assigned to the beginning of commit_lsn segment.
pub local_start_lsn: Option<Lsn>,
}

View File

@@ -19,8 +19,8 @@ thiserror = "1.0"
tokio = { version = "1.17", features = ["macros"]}
tokio-rustls = "0.23"
tracing = "0.1"
tracing-subscriber = { version = "0.3", features = ["env-filter"] }
nix = "0.23.0"
tracing-subscriber = { version = "0.3", features = ["env-filter", "json"] }
nix = "0.25"
signal-hook = "0.3.10"
rand = "0.8.3"
jsonwebtoken = "8"
@@ -28,8 +28,10 @@ hex = { version = "0.4.3", features = ["serde"] }
rustls = "0.20.2"
rustls-split = "0.3.0"
git-version = "0.3.5"
serde_with = "1.12.0"
serde_with = "2.0"
once_cell = "1.13.0"
strum = "0.24"
strum_macros = "0.24"
metrics = { path = "../metrics" }
@@ -40,7 +42,7 @@ byteorder = "1.4.3"
bytes = "1.0.1"
hex-literal = "0.3"
tempfile = "3.2"
criterion = "0.3"
criterion = "0.4"
rustls-pemfile = "1"
[[bench]]

View File

@@ -12,16 +12,8 @@ pub fn create_dir(path: impl AsRef<Path>) -> io::Result<()> {
let path = path.as_ref();
fs::create_dir(path)?;
File::open(path)?.sync_all()?;
if let Some(parent) = path.parent() {
File::open(parent)?.sync_all()
} else {
Err(io::Error::new(
io::ErrorKind::InvalidInput,
"can't find parent",
))
}
fsync_file_and_parent(path)?;
Ok(())
}
/// Similar to [`std::fs::create_dir_all`], except we fsync all
@@ -65,12 +57,12 @@ pub fn create_dir_all(path: impl AsRef<Path>) -> io::Result<()> {
// Fsync the created directories from child to parent.
for &path in dirs_to_create.iter() {
File::open(path)?.sync_all()?;
fsync(path)?;
}
// If we created any new directories, fsync the parent.
if !dirs_to_create.is_empty() {
File::open(path)?.sync_all()?;
fsync(path)?;
}
Ok(())
@@ -92,6 +84,33 @@ pub fn path_with_suffix_extension(original_path: impl AsRef<Path>, suffix: &str)
.with_extension(new_extension.as_ref())
}
pub fn fsync_file_and_parent(file_path: &Path) -> io::Result<()> {
let parent = file_path.parent().ok_or_else(|| {
io::Error::new(
io::ErrorKind::Other,
format!("File {file_path:?} has no parent"),
)
})?;
fsync(file_path)?;
fsync(parent)?;
Ok(())
}
pub fn fsync(path: &Path) -> io::Result<()> {
File::open(path)
.map_err(|e| io::Error::new(e.kind(), format!("Failed to open the file {path:?}: {e}")))
.and_then(|file| {
file.sync_all().map_err(|e| {
io::Error::new(
e.kind(),
format!("Failed to sync file {path:?} data and metadata: {e}"),
)
})
})
.map_err(|e| io::Error::new(e.kind(), format!("Failed to fsync file {path:?}: {e}")))
}
#[cfg(test)]
mod tests {
use tempfile::tempdir;

View File

@@ -9,6 +9,7 @@ use once_cell::sync::Lazy;
use routerify::ext::RequestExt;
use routerify::RequestInfo;
use routerify::{Middleware, Router, RouterBuilder, RouterService};
use tokio::task::JoinError;
use tracing::info;
use std::future::Future;
@@ -35,7 +36,13 @@ async fn prometheus_metrics_handler(_req: Request<Body>) -> Result<Response<Body
let mut buffer = vec![];
let encoder = TextEncoder::new();
let metrics = metrics::gather();
let metrics = tokio::task::spawn_blocking(move || {
// Currently we take a lot of mutexes while collecting metrics, so it's
// better to spawn a blocking task to avoid blocking the event loop.
metrics::gather()
})
.await
.map_err(|e: JoinError| ApiError::InternalServerError(e.into()))?;
encoder.encode(&metrics, &mut buffer).unwrap();
let response = Response::builder()

View File

@@ -75,6 +75,12 @@ impl From<[u8; 16]> for Id {
}
}
impl From<Id> for u128 {
fn from(id: Id) -> Self {
u128::from_le_bytes(id.0)
}
}
impl fmt::Display for Id {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(&self.hex_encode())
@@ -136,6 +142,12 @@ macro_rules! id_newtype {
}
}
impl From<$t> for u128 {
fn from(id: $t) -> Self {
u128::from(id.0)
}
}
impl fmt::Display for $t {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
self.0.fmt(f)

View File

@@ -22,8 +22,8 @@ pub mod pq_proto;
// dealing with connstring parsing and handy access to it's parts
pub mod connstring;
// helper functions for creating and fsyncing directories/trees
pub mod crashsafe_dir;
// helper functions for creating and fsyncing
pub mod crashsafe;
// common authentication routines
pub mod auth;

View File

@@ -1,11 +1,35 @@
use std::{
fs::{File, OpenOptions},
path::Path,
str::FromStr,
};
use anyhow::{Context, Result};
use strum_macros::{EnumString, EnumVariantNames};
pub fn init(log_filename: impl AsRef<Path>, daemonize: bool) -> Result<File> {
#[derive(EnumString, EnumVariantNames, Eq, PartialEq, Debug, Clone, Copy)]
#[strum(serialize_all = "snake_case")]
pub enum LogFormat {
Plain,
Json,
}
impl LogFormat {
pub fn from_config(s: &str) -> anyhow::Result<LogFormat> {
use strum::VariantNames;
LogFormat::from_str(s).with_context(|| {
format!(
"Unrecognized log format. Please specify one of: {:?}",
LogFormat::VARIANTS
)
})
}
}
pub fn init(
log_filename: impl AsRef<Path>,
daemonize: bool,
log_format: LogFormat,
) -> Result<File> {
// Don't open the same file for output multiple times;
// the different fds could overwrite each other's output.
let log_file = OpenOptions::new()
@@ -21,22 +45,50 @@ pub fn init(log_filename: impl AsRef<Path>, daemonize: bool) -> Result<File> {
let env_filter = tracing_subscriber::EnvFilter::try_from_default_env()
.unwrap_or_else(|_| tracing_subscriber::EnvFilter::new(default_filter_str));
let x: File = log_file.try_clone().unwrap();
let base_logger = tracing_subscriber::fmt()
.with_env_filter(env_filter)
.with_target(false) // don't include event targets
.with_ansi(false); // don't use colors in log file;
.with_target(false)
.with_ansi(false)
.with_writer(move || -> Box<dyn std::io::Write> {
// we are cloning and returning log file in order to allow redirecting daemonized stdout and stderr to it
// if we do not use daemonization (e.g. in docker) it is better to log to stdout directly
// for example to be in line with docker log command which expects logs comimg from stdout
if daemonize {
Box::new(x.try_clone().unwrap())
} else {
Box::new(std::io::stdout())
}
});
// we are cloning and returning log file in order to allow redirecting daemonized stdout and stderr to it
// if we do not use daemonization (e.g. in docker) it is better to log to stdout directly
// for example to be in line with docker log command which expects logs comimg from stdout
if daemonize {
let x = log_file.try_clone().unwrap();
base_logger
.with_writer(move || x.try_clone().unwrap())
.init();
} else {
base_logger.init();
match log_format {
LogFormat::Json => base_logger.json().init(),
LogFormat::Plain => base_logger.init(),
}
Ok(log_file)
}
// #[cfg(test)]
// Due to global logger, can't run tests in same process.
// So until there's a non-global one, the tests are in ../tests/ as separate files.
#[macro_export(local_inner_macros)]
macro_rules! test_init_file_logger {
($log_level:expr, $log_format:expr) => {{
use std::str::FromStr;
std::env::set_var("RUST_LOG", $log_level);
let tmp_dir = tempfile::TempDir::new().unwrap();
let log_file_path = tmp_dir.path().join("logfile");
let log_format = $crate::logging::LogFormat::from_str($log_format).unwrap();
let _log_file = $crate::logging::init(&log_file_path, true, log_format).unwrap();
let log_file = std::fs::OpenOptions::new()
.read(true)
.open(&log_file_path)
.unwrap();
log_file
}};
}

View File

@@ -66,6 +66,11 @@ impl Lsn {
(self.0 % seg_sz as u64) as usize
}
/// Compute LSN of the segment start.
pub fn segment_lsn(self, seg_sz: usize) -> Lsn {
Lsn(self.0 - (self.0 % seg_sz as u64))
}
/// Compute the segment number
pub fn segment_number(self, seg_sz: usize) -> u64 {
self.0 / seg_sz as u64

View File

@@ -15,7 +15,7 @@ use std::sync::Arc;
use std::task::Poll;
use tracing::{debug, error, trace};
use tokio::io::{AsyncRead, AsyncWrite, AsyncWriteExt};
use tokio::io::{AsyncRead, AsyncWrite, AsyncWriteExt, BufReader};
use tokio_rustls::TlsAcceptor;
#[async_trait::async_trait]
@@ -66,8 +66,8 @@ pub enum ProcessMsgResult {
/// Always-writeable sock_split stream.
/// May not be readable. See [`PostgresBackend::take_stream_in`]
pub enum Stream {
Unencrypted(tokio::net::TcpStream),
Tls(Box<tokio_rustls::server::TlsStream<tokio::net::TcpStream>>),
Unencrypted(BufReader<tokio::net::TcpStream>),
Tls(Box<tokio_rustls::server::TlsStream<BufReader<tokio::net::TcpStream>>>),
Broken,
}
@@ -157,7 +157,7 @@ impl PostgresBackend {
let peer_addr = socket.peer_addr()?;
Ok(Self {
stream: Stream::Unencrypted(socket),
stream: Stream::Unencrypted(BufReader::new(socket)),
buf_out: BytesMut::with_capacity(10 * 1024),
state: ProtoState::Initialization,
md5_salt: [0u8; 4],

View File

@@ -10,6 +10,7 @@ use serde::{Deserialize, Serialize};
use std::{
borrow::Cow,
collections::HashMap,
fmt,
future::Future,
io::{self, Cursor},
str,
@@ -124,6 +125,19 @@ pub struct CancelKeyData {
pub cancel_key: i32,
}
impl fmt::Display for CancelKeyData {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
let hi = (self.backend_pid as u64) << 32;
let lo = self.cancel_key as u64;
let id = hi | lo;
// This format is more compact and might work better for logs.
f.debug_tuple("CancelKeyData")
.field(&format_args!("{:x}", id))
.finish()
}
}
use rand::distributions::{Distribution, Standard};
impl Distribution<CancelKeyData> for Standard {
fn sample<R: rand::Rng + ?Sized>(&self, rng: &mut R) -> CancelKeyData {

View File

@@ -0,0 +1,36 @@
// This could be in ../src/logging.rs but since the logger is global, these
// can't be run in threads of the same process
use std::fs::File;
use std::io::{BufRead, BufReader, Lines};
use tracing::*;
use utils::test_init_file_logger;
fn read_lines(file: File) -> Lines<BufReader<File>> {
BufReader::new(file).lines()
}
#[test]
fn test_json_format_has_message_and_custom_field() {
std::env::set_var("RUST_LOG", "info");
let log_file = test_init_file_logger!("info", "json");
let custom_field: &str = "hi";
trace!(custom = %custom_field, "test log message");
debug!(custom = %custom_field, "test log message");
info!(custom = %custom_field, "test log message");
warn!(custom = %custom_field, "test log message");
error!(custom = %custom_field, "test log message");
let lines = read_lines(log_file);
for line in lines {
let content = line.unwrap();
let json_object = serde_json::from_str::<serde_json::Value>(&content).unwrap();
assert_eq!(json_object["fields"]["custom"], "hi");
assert_eq!(json_object["fields"]["message"], "test log message");
assert_ne!(json_object["level"], "TRACE");
assert_ne!(json_object["level"], "DEBUG");
}
}

View File

@@ -0,0 +1,36 @@
// This could be in ../src/logging.rs but since the logger is global, these
// can't be run in threads of the same process
use std::fs::File;
use std::io::{BufRead, BufReader, Lines};
use tracing::*;
use utils::test_init_file_logger;
fn read_lines(file: File) -> Lines<BufReader<File>> {
BufReader::new(file).lines()
}
#[test]
fn test_plain_format_has_message_and_custom_field() {
std::env::set_var("RUST_LOG", "warn");
let log_file = test_init_file_logger!("warn", "plain");
let custom_field: &str = "hi";
trace!(custom = %custom_field, "test log message");
debug!(custom = %custom_field, "test log message");
info!(custom = %custom_field, "test log message");
warn!(custom = %custom_field, "test log message");
error!(custom = %custom_field, "test log message");
let lines = read_lines(log_file);
for line in lines {
let content = line.unwrap();
serde_json::from_str::<serde_json::Value>(&content).unwrap_err();
assert!(content.contains("custom=hi"));
assert!(content.contains("test log message"));
assert!(!content.contains("TRACE"));
assert!(!content.contains("DEBUG"));
assert!(!content.contains("INFO"));
}
}

View File

@@ -23,7 +23,7 @@ futures = "0.3.13"
hex = "0.4.3"
hyper = "0.14"
itertools = "0.10.3"
clap = "3.0"
clap = { version = "4.0", features = ["string"] }
daemonize = "0.4.1"
tokio = { version = "1.17", features = ["process", "sync", "macros", "fs", "rt", "io-util", "time"] }
tokio-util = { version = "0.7.3", features = ["io", "io-util"] }
@@ -38,26 +38,27 @@ tar = "0.4.33"
humantime = "2.1.0"
serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
serde_with = "1.12.0"
serde_with = "2.0"
humantime-serde = "1.1.1"
pprof = { git = "https://github.com/neondatabase/pprof-rs.git", branch = "wallclock-profiling", features = ["flamegraph"], optional = true }
toml_edit = { version = "0.13", features = ["easy"] }
toml_edit = { version = "0.14", features = ["easy"] }
scopeguard = "1.1.0"
const_format = "0.2.21"
tracing = "0.1.36"
signal-hook = "0.3.10"
url = "2"
nix = "0.23"
nix = "0.25"
once_cell = "1.13.0"
crossbeam-utils = "0.8.5"
fail = "0.5.0"
git-version = "0.3.5"
rstar = "0.9.3"
num-traits = "0.2.15"
amplify_num = "0.4.1"
amplify_num = { git = "https://github.com/hlinnaka/rust-amplify.git", branch = "unsigned-int-perf" }
pageserver_api = { path = "../libs/pageserver_api" }
postgres_ffi = { path = "../libs/postgres_ffi" }
etcd_broker = { path = "../libs/etcd_broker" }
metrics = { path = "../libs/metrics" }
@@ -66,7 +67,13 @@ remote_storage = { path = "../libs/remote_storage" }
workspace_hack = { version = "0.1", path = "../workspace_hack" }
close_fds = "0.3.2"
walkdir = "2.3.2"
svg_fmt = "0.4.1"
[dev-dependencies]
criterion = "0.4"
hex-literal = "0.3"
tempfile = "3.2"
[[bench]]
name = "bench_layer_map"
harness = false

File diff suppressed because it is too large Load Diff

View File

@@ -22,8 +22,8 @@ use std::time::SystemTime;
use tar::{Builder, EntryType, Header};
use tracing::*;
use crate::reltag::{RelTag, SlruKind};
use crate::tenant::Timeline;
use pageserver_api::reltag::{RelTag, SlruKind};
use postgres_ffi::pg_constants::{DEFAULTTABLESPACE_OID, GLOBALTABLESPACE_OID};
use postgres_ffi::pg_constants::{PGDATA_SPECIAL_FILES, PGDATA_SUBDIRS, PG_HBA};

View File

@@ -0,0 +1,150 @@
//! A tool for visualizing the arrangement of layerfiles within a timeline.
//!
//! It reads filenames from stdin and prints a svg on stdout. The image is a plot in
//! page-lsn space, where every delta layer is a rectangle and every image layer is a
//! thick line. Legend:
//! - The x axis (left to right) represents page index.
//! - The y axis represents LSN, growing upwards.
//!
//! Coordinates in both axis are compressed for better readability.
//! (see https://medium.com/algorithms-digest/coordinate-compression-2fff95326fb)
//!
//! Example use:
//! ```
//! $ cd test_output/test_pgbench\[neon-45-684\]/repo/tenants/$TENANT/timelines/$TIMELINE
//! $ ls | grep "__" | cargo run --release --bin draw_timeline_dir > out.svg
//! $ firefox out.svg
//! ```
//!
//! This API was chosen so that we can easily work with filenames extracted from ssh,
//! or from pageserver log files.
//!
//! TODO Consider shipping this as a grafana panel plugin:
//! https://grafana.com/tutorials/build-a-panel-plugin/
use anyhow::Result;
use pageserver::repository::Key;
use std::cmp::Ordering;
use std::io::{self, BufRead};
use std::{
collections::{BTreeMap, BTreeSet},
ops::Range,
};
use svg_fmt::{rectangle, rgb, BeginSvg, EndSvg, Fill, Stroke};
use utils::{lsn::Lsn, project_git_version};
project_git_version!(GIT_VERSION);
// Map values to their compressed coordinate - the index the value
// would have in a sorted and deduplicated list of all values.
fn build_coordinate_compression_map<T: Ord + Copy>(coords: Vec<T>) -> BTreeMap<T, usize> {
let set: BTreeSet<T> = coords.into_iter().collect();
let mut map: BTreeMap<T, usize> = BTreeMap::new();
for (i, e) in set.iter().enumerate() {
map.insert(*e, i);
}
map
}
fn parse_filename(name: &str) -> (Range<Key>, Range<Lsn>) {
let split: Vec<&str> = name.split("__").collect();
let keys: Vec<&str> = split[0].split('-').collect();
let mut lsns: Vec<&str> = split[1].split('-').collect();
if lsns.len() == 1 {
lsns.push(lsns[0]);
}
let keys = Key::from_hex(keys[0]).unwrap()..Key::from_hex(keys[1]).unwrap();
let lsns = Lsn::from_hex(lsns[0]).unwrap()..Lsn::from_hex(lsns[1]).unwrap();
(keys, lsns)
}
fn main() -> Result<()> {
// Parse layer filenames from stdin
let mut ranges: Vec<(Range<Key>, Range<Lsn>)> = vec![];
let stdin = io::stdin();
for line in stdin.lock().lines() {
let range = parse_filename(&line.unwrap());
ranges.push(range);
}
// Collect all coordinates
let mut keys: Vec<Key> = vec![];
let mut lsns: Vec<Lsn> = vec![];
for (keyr, lsnr) in &ranges {
keys.push(keyr.start);
keys.push(keyr.end);
lsns.push(lsnr.start);
lsns.push(lsnr.end);
}
// Analyze
let key_map = build_coordinate_compression_map(keys);
let lsn_map = build_coordinate_compression_map(lsns);
// Initialize stats
let mut num_deltas = 0;
let mut num_images = 0;
// Draw
let stretch = 3.0; // Stretch out vertically for better visibility
println!(
"{}",
BeginSvg {
w: key_map.len() as f32,
h: stretch * lsn_map.len() as f32
}
);
for (keyr, lsnr) in &ranges {
let key_start = *key_map.get(&keyr.start).unwrap();
let key_end = *key_map.get(&keyr.end).unwrap();
let key_diff = key_end - key_start;
let lsn_max = lsn_map.len();
if key_start >= key_end {
panic!("Invalid key range {}-{}", key_start, key_end);
}
let lsn_start = *lsn_map.get(&lsnr.start).unwrap();
let lsn_end = *lsn_map.get(&lsnr.end).unwrap();
let mut lsn_diff = (lsn_end - lsn_start) as f32;
let mut fill = Fill::None;
let mut margin = 0.05 * lsn_diff; // Height-dependent margin to disambiguate overlapping deltas
let mut lsn_offset = 0.0;
// Fill in and thicken rectangle if it's an
// image layer so that we can see it.
match lsn_start.cmp(&lsn_end) {
Ordering::Less => num_deltas += 1,
Ordering::Equal => {
num_images += 1;
lsn_diff = 0.3;
lsn_offset = -lsn_diff / 2.0;
margin = 0.05;
fill = Fill::Color(rgb(0, 0, 0));
}
Ordering::Greater => panic!("Invalid lsn range {}-{}", lsn_start, lsn_end),
}
println!(
" {}",
rectangle(
key_start as f32 + stretch * margin,
stretch * (lsn_max as f32 - (lsn_end as f32 - margin - lsn_offset)),
key_diff as f32 - stretch * 2.0 * margin,
stretch * (lsn_diff - 2.0 * margin)
)
.fill(fill)
.stroke(Stroke::Color(rgb(0, 0, 0), 0.1))
.border_radius(0.4)
);
}
println!("{}", EndSvg);
eprintln!("num_images: {}", num_images);
eprintln!("num_deltas: {}", num_deltas);
Ok(())
}

View File

@@ -1,35 +0,0 @@
//! Main entry point for the dump_layerfile executable
//!
//! A handy tool for debugging, that's all.
use anyhow::Result;
use clap::{App, Arg};
use pageserver::page_cache;
use pageserver::tenant::dump_layerfile_from_path;
use pageserver::virtual_file;
use std::path::PathBuf;
use utils::project_git_version;
project_git_version!(GIT_VERSION);
fn main() -> Result<()> {
let arg_matches = App::new("Neon dump_layerfile utility")
.about("Dump contents of one layer file, for debugging")
.version(GIT_VERSION)
.arg(
Arg::new("path")
.help("Path to file to dump")
.required(true)
.index(1),
)
.get_matches();
let path = PathBuf::from(arg_matches.value_of("path").unwrap());
// Basic initialization of things that don't change after startup
virtual_file::init(10);
page_cache::init(100);
dump_layerfile_from_path(&path, true)?;
Ok(())
}

View File

@@ -6,13 +6,15 @@ use tracing::*;
use anyhow::{anyhow, bail, Context, Result};
use clap::{App, Arg};
use clap::{Arg, ArgAction, Command};
use daemonize::Daemonize;
use fail::FailScenario;
use metrics::set_build_info_metric;
use pageserver::{
config::{defaults::*, PageServerConf},
http, page_cache, page_service, profiling, task_mgr,
http, page_cache, page_image_cache, page_service, profiling, task_mgr,
task_mgr::TaskKind,
task_mgr::{
BACKGROUND_RUNTIME, COMPUTE_REQUEST_RUNTIME, MGMT_REQUEST_RUNTIME, WALRECEIVER_RUNTIME,
@@ -31,72 +33,35 @@ use utils::{
project_git_version!(GIT_VERSION);
const FEATURES: &[&str] = &[
#[cfg(feature = "testing")]
"testing",
#[cfg(feature = "fail/failpoints")]
"fail/failpoints",
#[cfg(feature = "profiling")]
"profiling",
];
fn version() -> String {
format!(
"{GIT_VERSION} profiling:{} failpoints:{}",
cfg!(feature = "profiling"),
fail::has_failpoints()
"{GIT_VERSION} failpoints: {}, features: {:?}",
fail::has_failpoints(),
FEATURES,
)
}
fn main() -> anyhow::Result<()> {
let arg_matches = App::new("Neon page server")
.about("Materializes WAL stream to pages and serves them to the postgres")
.version(&*version())
.arg(
let arg_matches = cli().get_matches();
Arg::new("daemonize")
.short('d')
.long("daemonize")
.takes_value(false)
.help("Run in the background"),
)
.arg(
Arg::new("init")
.long("init")
.takes_value(false)
.help("Initialize pageserver with all given config overrides"),
)
.arg(
Arg::new("workdir")
.short('D')
.long("workdir")
.takes_value(true)
.help("Working directory for the pageserver"),
)
// See `settings.md` for more details on the extra configuration patameters pageserver can process
.arg(
Arg::new("config-override")
.short('c')
.takes_value(true)
.number_of_values(1)
.multiple_occurrences(true)
.help("Additional configuration overrides of the ones from the toml config file (or new ones to add there).
Any option has to be a valid toml document, example: `-c=\"foo='hey'\"` `-c=\"foo={value=1}\"`"),
)
.arg(Arg::new("update-config").long("update-config").takes_value(false).help(
"Update the config file when started",
))
.arg(
Arg::new("enabled-features")
.long("enabled-features")
.takes_value(false)
.help("Show enabled compile time features"),
)
.get_matches();
if arg_matches.is_present("enabled-features") {
let features: &[&str] = &[
#[cfg(feature = "testing")]
"testing",
#[cfg(feature = "profiling")]
"profiling",
];
println!("{{\"features\": {features:?} }}");
if arg_matches.get_flag("enabled-features") {
println!("{{\"features\": {FEATURES:?} }}");
return Ok(());
}
let workdir = Path::new(arg_matches.value_of("workdir").unwrap_or(".neon"));
let workdir = arg_matches
.get_one::<String>("workdir")
.map(Path::new)
.unwrap_or_else(|| Path::new(".neon"));
let workdir = workdir
.canonicalize()
.with_context(|| format!("Error opening workdir '{}'", workdir.display()))?;
@@ -110,7 +75,7 @@ fn main() -> anyhow::Result<()> {
)
})?;
let daemonize = arg_matches.is_present("daemonize");
let daemonize = arg_matches.get_flag("daemonize");
let conf = match initialize_config(&cfg_file_path, arg_matches, &workdir)? {
ControlFlow::Continue(conf) => conf,
@@ -122,7 +87,7 @@ fn main() -> anyhow::Result<()> {
let tenants_path = conf.tenants_path();
if !tenants_path.exists() {
utils::crashsafe_dir::create_dir_all(conf.tenants_path()).with_context(|| {
utils::crashsafe::create_dir_all(conf.tenants_path()).with_context(|| {
format!(
"Failed to create tenants root dir at '{}'",
tenants_path.display()
@@ -136,6 +101,7 @@ fn main() -> anyhow::Result<()> {
// Basic initialization of things that don't change after startup
virtual_file::init(conf.max_file_descriptors);
page_cache::init(conf.page_cache_size);
page_image_cache::init(64 * conf.page_cache_size); // temporary hack for benchmarking
start_pageserver(conf, daemonize).context("Failed to start pageserver")?;
@@ -148,8 +114,8 @@ fn initialize_config(
arg_matches: clap::ArgMatches,
workdir: &Path,
) -> anyhow::Result<ControlFlow<(), &'static PageServerConf>> {
let init = arg_matches.is_present("init");
let update_config = init || arg_matches.is_present("update-config");
let init = arg_matches.get_flag("init");
let update_config = init || arg_matches.get_flag("update-config");
let (mut toml, config_file_exists) = if cfg_file_path.is_file() {
if init {
@@ -191,13 +157,10 @@ fn initialize_config(
)
};
if let Some(values) = arg_matches.values_of("config-override") {
if let Some(values) = arg_matches.get_many::<String>("config-override") {
for option_line in values {
let doc = toml_edit::Document::from_str(option_line).with_context(|| {
format!(
"Option '{}' could not be parsed as a toml document",
option_line
)
format!("Option '{option_line}' could not be parsed as a toml document")
})?;
for (key, item) in doc.iter() {
@@ -237,9 +200,9 @@ fn initialize_config(
fn start_pageserver(conf: &'static PageServerConf, daemonize: bool) -> Result<()> {
// Initialize logger
let log_file = logging::init(LOG_FILE_NAME, daemonize)?;
let log_file = logging::init(LOG_FILE_NAME, daemonize, conf.log_format)?;
info!("version: {GIT_VERSION}");
info!("version: {}", version());
// TODO: Check that it looks like a valid repository before going further
@@ -356,6 +319,8 @@ fn start_pageserver(conf: &'static PageServerConf, daemonize: bool) -> Result<()
},
);
set_build_info_metric(GIT_VERSION);
// All started up! Now just sit and wait for shutdown signal.
signals.handle(|signal| match signal {
Signal::Quit => {
@@ -378,3 +343,55 @@ fn start_pageserver(conf: &'static PageServerConf, daemonize: bool) -> Result<()
}
})
}
fn cli() -> Command {
Command::new("Neon page server")
.about("Materializes WAL stream to pages and serves them to the postgres")
.version(version())
.arg(
Arg::new("daemonize")
.short('d')
.long("daemonize")
.action(ArgAction::SetTrue)
.help("Run in the background"),
)
.arg(
Arg::new("init")
.long("init")
.action(ArgAction::SetTrue)
.help("Initialize pageserver with all given config overrides"),
)
.arg(
Arg::new("workdir")
.short('D')
.long("workdir")
.help("Working directory for the pageserver"),
)
// See `settings.md` for more details on the extra configuration patameters pageserver can process
.arg(
Arg::new("config-override")
.short('c')
.num_args(1)
.action(ArgAction::Append)
.help("Additional configuration overrides of the ones from the toml config file (or new ones to add there). \
Any option has to be a valid toml document, example: `-c=\"foo='hey'\"` `-c=\"foo={value=1}\"`"),
)
.arg(
Arg::new("update-config")
.long("update-config")
.action(ArgAction::SetTrue)
.help("Update the config file when started"),
)
.arg(
Arg::new("enabled-features")
.long("enabled-features")
.action(ArgAction::SetTrue)
.help("Show enabled compile time features"),
)
}
#[test]
fn verify_cli() {
cli().debug_assert();
}

View File

@@ -0,0 +1,154 @@
//! A helper tool to manage pageserver binary files.
//! Accepts a file as an argument, attempts to parse it with all ways possible
//! and prints its interpreted context.
//!
//! Separate, `metadata` subcommand allows to print and update pageserver's metadata file.
use std::{
path::{Path, PathBuf},
str::FromStr,
};
use anyhow::Context;
use clap::{value_parser, Arg, Command};
use pageserver::{
page_cache,
tenant::{dump_layerfile_from_path, metadata::TimelineMetadata},
virtual_file,
};
use postgres_ffi::ControlFileData;
use utils::{lsn::Lsn, project_git_version};
project_git_version!(GIT_VERSION);
const METADATA_SUBCOMMAND: &str = "metadata";
fn main() -> anyhow::Result<()> {
let arg_matches = cli().get_matches();
match arg_matches.subcommand() {
Some((subcommand_name, subcommand_matches)) => {
let path = subcommand_matches
.get_one::<PathBuf>("metadata_path")
.context("'metadata_path' argument is missing")?
.to_path_buf();
anyhow::ensure!(
subcommand_name == METADATA_SUBCOMMAND,
"Unknown subcommand {subcommand_name}"
);
handle_metadata(&path, subcommand_matches)?;
}
None => {
let path = arg_matches
.get_one::<PathBuf>("path")
.context("'path' argument is missing")?
.to_path_buf();
println!(
"No subcommand specified, attempting to guess the format for file {}",
path.display()
);
if let Err(e) = read_pg_control_file(&path) {
println!(
"Failed to read input file as a pg control one: {e:#}\n\
Attempting to read it as layer file"
);
print_layerfile(&path)?;
}
}
};
Ok(())
}
fn read_pg_control_file(control_file_path: &Path) -> anyhow::Result<()> {
let control_file = ControlFileData::decode(&std::fs::read(&control_file_path)?)?;
println!("{control_file:?}");
let control_file_initdb = Lsn(control_file.checkPoint);
println!(
"pg_initdb_lsn: {}, aligned: {}",
control_file_initdb,
control_file_initdb.align()
);
Ok(())
}
fn print_layerfile(path: &Path) -> anyhow::Result<()> {
// Basic initialization of things that don't change after startup
virtual_file::init(10);
page_cache::init(100);
dump_layerfile_from_path(path, true)
}
fn handle_metadata(path: &Path, arg_matches: &clap::ArgMatches) -> Result<(), anyhow::Error> {
let metadata_bytes = std::fs::read(&path)?;
let mut meta = TimelineMetadata::from_bytes(&metadata_bytes)?;
println!("Current metadata:\n{meta:?}");
let mut update_meta = false;
if let Some(disk_consistent_lsn) = arg_matches.get_one::<String>("disk_consistent_lsn") {
meta = TimelineMetadata::new(
Lsn::from_str(disk_consistent_lsn)?,
meta.prev_record_lsn(),
meta.ancestor_timeline(),
meta.ancestor_lsn(),
meta.latest_gc_cutoff_lsn(),
meta.initdb_lsn(),
meta.pg_version(),
);
update_meta = true;
}
if let Some(prev_record_lsn) = arg_matches.get_one::<String>("prev_record_lsn") {
meta = TimelineMetadata::new(
meta.disk_consistent_lsn(),
Some(Lsn::from_str(prev_record_lsn)?),
meta.ancestor_timeline(),
meta.ancestor_lsn(),
meta.latest_gc_cutoff_lsn(),
meta.initdb_lsn(),
meta.pg_version(),
);
update_meta = true;
}
if update_meta {
let metadata_bytes = meta.to_bytes()?;
std::fs::write(&path, &metadata_bytes)?;
}
Ok(())
}
fn cli() -> Command {
Command::new("Neon Pageserver binutils")
.about("Reads pageserver (and related) binary files management utility")
.version(GIT_VERSION)
.arg(
Arg::new("path")
.help("Input file path")
.value_parser(value_parser!(PathBuf))
.required(false),
)
.subcommand(
Command::new(METADATA_SUBCOMMAND)
.about("Read and update pageserver metadata file")
.arg(
Arg::new("metadata_path")
.help("Input metadata file path")
.value_parser(value_parser!(PathBuf))
.required(false),
)
.arg(
Arg::new("disk_consistent_lsn")
.long("disk_consistent_lsn")
.help("Replace disk consistent Lsn"),
)
.arg(
Arg::new("prev_record_lsn")
.long("prev_record_lsn")
.help("Replace previous record Lsn"),
),
)
}
#[test]
fn verify_cli() {
cli().debug_assert();
}

View File

@@ -1,75 +0,0 @@
//! Main entry point for the edit_metadata executable
//!
//! A handy tool for debugging, that's all.
use anyhow::Result;
use clap::{App, Arg};
use pageserver::tenant::metadata::TimelineMetadata;
use std::path::PathBuf;
use std::str::FromStr;
use utils::{lsn::Lsn, project_git_version};
project_git_version!(GIT_VERSION);
fn main() -> Result<()> {
let arg_matches = App::new("Neon update metadata utility")
.about("Dump or update metadata file")
.version(GIT_VERSION)
.arg(
Arg::new("path")
.help("Path to metadata file")
.required(true),
)
.arg(
Arg::new("disk_lsn")
.short('d')
.long("disk_lsn")
.takes_value(true)
.help("Replace disk constistent lsn"),
)
.arg(
Arg::new("prev_lsn")
.short('p')
.long("prev_lsn")
.takes_value(true)
.help("Previous record LSN"),
)
.get_matches();
let path = PathBuf::from(arg_matches.value_of("path").unwrap());
let metadata_bytes = std::fs::read(&path)?;
let mut meta = TimelineMetadata::from_bytes(&metadata_bytes)?;
println!("Current metadata:\n{:?}", &meta);
let mut update_meta = false;
if let Some(disk_lsn) = arg_matches.value_of("disk_lsn") {
meta = TimelineMetadata::new(
Lsn::from_str(disk_lsn)?,
meta.prev_record_lsn(),
meta.ancestor_timeline(),
meta.ancestor_lsn(),
meta.latest_gc_cutoff_lsn(),
meta.initdb_lsn(),
meta.pg_version(),
);
update_meta = true;
}
if let Some(prev_lsn) = arg_matches.value_of("prev_lsn") {
meta = TimelineMetadata::new(
meta.disk_consistent_lsn(),
Some(Lsn::from_str(prev_lsn)?),
meta.ancestor_timeline(),
meta.ancestor_lsn(),
meta.latest_gc_cutoff_lsn(),
meta.initdb_lsn(),
meta.pg_version(),
);
update_meta = true;
}
if update_meta {
let metadata_bytes = meta.to_bytes()?;
std::fs::write(&path, &metadata_bytes)?;
}
Ok(())
}

View File

@@ -7,6 +7,7 @@
use anyhow::{anyhow, bail, ensure, Context, Result};
use remote_storage::RemoteStorageConfig;
use std::env;
use utils::crashsafe::path_with_suffix_extension;
use std::path::{Path, PathBuf};
use std::str::FromStr;
@@ -16,6 +17,7 @@ use toml_edit::{Document, Item};
use url::Url;
use utils::{
id::{NodeId, TenantId, TimelineId},
logging::LogFormat,
postgres_backend::AuthType,
};
@@ -24,16 +26,17 @@ use crate::tenant_config::{TenantConf, TenantConfOpt};
/// The name of the metadata file pageserver creates per timeline.
pub const METADATA_FILE_NAME: &str = "metadata";
pub const TIMELINE_UNINIT_MARK_SUFFIX: &str = "___uninit";
const TENANT_CONFIG_NAME: &str = "config";
pub mod defaults {
use crate::tenant_config::defaults::*;
use const_format::formatcp;
pub const DEFAULT_PG_LISTEN_PORT: u16 = 64000;
pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 9898;
pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");
pub use pageserver_api::{
DEFAULT_HTTP_LISTEN_ADDR, DEFAULT_HTTP_LISTEN_PORT, DEFAULT_PG_LISTEN_ADDR,
DEFAULT_PG_LISTEN_PORT,
};
pub const DEFAULT_WAIT_LSN_TIMEOUT: &str = "60 s";
pub const DEFAULT_WAL_REDO_TIMEOUT: &str = "60 s";
@@ -43,6 +46,8 @@ pub mod defaults {
pub const DEFAULT_PAGE_CACHE_SIZE: usize = 8192;
pub const DEFAULT_MAX_FILE_DESCRIPTORS: usize = 100;
pub const DEFAULT_LOG_FORMAT: &str = "plain";
///
/// Default built-in configuration file.
///
@@ -61,6 +66,7 @@ pub mod defaults {
# initial superuser role name to use when creating a new tenant
#initial_superuser_name = '{DEFAULT_SUPERUSER}'
#log_format = '{DEFAULT_LOG_FORMAT}'
# [tenant_config]
#checkpoint_distance = {DEFAULT_CHECKPOINT_DISTANCE} # in bytes
#checkpoint_timeout = {DEFAULT_CHECKPOINT_TIMEOUT}
@@ -124,6 +130,8 @@ pub struct PageServerConf {
/// Etcd broker endpoints to connect to.
pub broker_endpoints: Vec<Url>,
pub log_format: LogFormat,
}
#[derive(Debug, Clone, PartialEq, Eq)]
@@ -190,6 +198,8 @@ struct PageServerConfigBuilder {
profiling: BuilderValue<ProfilingConfig>,
broker_etcd_prefix: BuilderValue<String>,
broker_endpoints: BuilderValue<Vec<Url>>,
log_format: BuilderValue<LogFormat>,
}
impl Default for PageServerConfigBuilder {
@@ -217,6 +227,7 @@ impl Default for PageServerConfigBuilder {
profiling: Set(ProfilingConfig::Disabled),
broker_etcd_prefix: Set(etcd_broker::DEFAULT_NEON_BROKER_ETCD_PREFIX.to_string()),
broker_endpoints: Set(Vec::new()),
log_format: Set(LogFormat::from_str(DEFAULT_LOG_FORMAT).unwrap()),
}
}
}
@@ -289,6 +300,10 @@ impl PageServerConfigBuilder {
self.profiling = BuilderValue::Set(profiling)
}
pub fn log_format(&mut self, log_format: LogFormat) {
self.log_format = BuilderValue::Set(log_format)
}
pub fn build(self) -> anyhow::Result<PageServerConf> {
let broker_endpoints = self
.broker_endpoints
@@ -333,6 +348,7 @@ impl PageServerConfigBuilder {
broker_etcd_prefix: self
.broker_etcd_prefix
.ok_or(anyhow!("missing broker_etcd_prefix"))?,
log_format: self.log_format.ok_or(anyhow!("missing log_format"))?,
})
}
}
@@ -364,6 +380,17 @@ impl PageServerConf {
self.timelines_path(tenant_id).join(timeline_id.to_string())
}
pub fn timeline_uninit_mark_file_path(
&self,
tenant_id: TenantId,
timeline_id: TimelineId,
) -> PathBuf {
path_with_suffix_extension(
self.timeline_path(&timeline_id, &tenant_id),
TIMELINE_UNINIT_MARK_SUFFIX,
)
}
/// Points to a place in pageserver's local directory,
/// where certain timeline's metadata file should be located.
pub fn metadata_path(&self, timeline_id: TimelineId, tenant_id: TenantId) -> PathBuf {
@@ -374,28 +401,28 @@ impl PageServerConf {
//
// Postgres distribution paths
//
pub fn pg_distrib_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_distrib_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
let path = self.pg_distrib_dir.clone();
match pg_version {
14 => path.join(format!("v{pg_version}")),
15 => path.join(format!("v{pg_version}")),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(path.join(format!("v{pg_version}"))),
15 => Ok(path.join(format!("v{pg_version}"))),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_bin_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_bin_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("bin"),
15 => self.pg_distrib_dir(pg_version).join("bin"),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(self.pg_distrib_dir(pg_version)?.join("bin")),
15 => Ok(self.pg_distrib_dir(pg_version)?.join("bin")),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_lib_dir(&self, pg_version: u32) -> PathBuf {
pub fn pg_lib_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("lib"),
15 => self.pg_distrib_dir(pg_version).join("lib"),
_ => panic!("Unsupported postgres version: {}", pg_version),
14 => Ok(self.pg_distrib_dir(pg_version)?.join("lib")),
15 => Ok(self.pg_distrib_dir(pg_version)?.join("lib")),
_ => bail!("Unsupported postgres version: {}", pg_version),
}
}
@@ -446,6 +473,9 @@ impl PageServerConf {
})
.collect::<anyhow::Result<_>>()?,
),
"log_format" => builder.log_format(
LogFormat::from_config(&parse_toml_string(key, item)?)?
),
_ => bail!("unrecognized pageserver option '{key}'"),
}
}
@@ -558,6 +588,7 @@ impl PageServerConf {
default_tenant_conf: TenantConf::dummy_conf(),
broker_endpoints: Vec::new(),
broker_etcd_prefix: etcd_broker::DEFAULT_NEON_BROKER_ETCD_PREFIX.to_string(),
log_format: LogFormat::from_str(defaults::DEFAULT_LOG_FORMAT).unwrap(),
}
}
}
@@ -652,6 +683,8 @@ max_file_descriptors = 333
initial_superuser_name = 'zzzz'
id = 10
log_format = 'json'
"#;
#[test]
@@ -691,6 +724,7 @@ id = 10
.parse()
.expect("Failed to parse a valid broker endpoint URL")],
broker_etcd_prefix: etcd_broker::DEFAULT_NEON_BROKER_ETCD_PREFIX.to_string(),
log_format: LogFormat::from_str(defaults::DEFAULT_LOG_FORMAT).unwrap(),
},
"Correct defaults should be used when no config values are provided"
);
@@ -735,6 +769,7 @@ id = 10
.parse()
.expect("Failed to parse a valid broker endpoint URL")],
broker_etcd_prefix: etcd_broker::DEFAULT_NEON_BROKER_ETCD_PREFIX.to_string(),
log_format: LogFormat::Json,
},
"Should be able to parse all basic config values correctly"
);

View File

@@ -1,3 +1,4 @@
pub mod models;
pub mod routes;
pub use routes::make_router;
pub use pageserver_api::models;

View File

@@ -1,182 +0,0 @@
use std::num::NonZeroU64;
use serde::{Deserialize, Serialize};
use serde_with::{serde_as, DisplayFromStr};
use utils::{
id::{NodeId, TenantId, TimelineId},
lsn::Lsn,
};
use crate::tenant::TenantState;
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TimelineCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_start_lsn: Option<Lsn>,
pub pg_version: Option<u32>,
}
#[serde_as]
#[derive(Serialize, Deserialize, Default)]
pub struct TenantCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_tenant_id: Option<TenantId>,
pub checkpoint_distance: Option<u64>,
pub checkpoint_timeout: Option<String>,
pub compaction_target_size: Option<u64>,
pub compaction_period: Option<String>,
pub compaction_threshold: Option<usize>,
pub gc_horizon: Option<u64>,
pub gc_period: Option<String>,
pub image_creation_threshold: Option<usize>,
pub pitr_interval: Option<String>,
pub walreceiver_connect_timeout: Option<String>,
pub lagging_wal_timeout: Option<String>,
pub max_lsn_wal_lag: Option<NonZeroU64>,
}
#[serde_as]
#[derive(Serialize, Deserialize)]
#[serde(transparent)]
pub struct TenantCreateResponse(#[serde_as(as = "DisplayFromStr")] pub TenantId);
#[derive(Serialize)]
pub struct StatusResponse {
pub id: NodeId,
}
impl TenantCreateRequest {
pub fn new(new_tenant_id: Option<TenantId>) -> TenantCreateRequest {
TenantCreateRequest {
new_tenant_id,
..Default::default()
}
}
}
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TenantConfigRequest {
pub tenant_id: TenantId,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub checkpoint_distance: Option<u64>,
pub checkpoint_timeout: Option<String>,
pub compaction_target_size: Option<u64>,
pub compaction_period: Option<String>,
pub compaction_threshold: Option<usize>,
pub gc_horizon: Option<u64>,
pub gc_period: Option<String>,
pub image_creation_threshold: Option<usize>,
pub pitr_interval: Option<String>,
pub walreceiver_connect_timeout: Option<String>,
pub lagging_wal_timeout: Option<String>,
pub max_lsn_wal_lag: Option<NonZeroU64>,
}
impl TenantConfigRequest {
pub fn new(tenant_id: TenantId) -> TenantConfigRequest {
TenantConfigRequest {
tenant_id,
checkpoint_distance: None,
checkpoint_timeout: None,
compaction_target_size: None,
compaction_period: None,
compaction_threshold: None,
gc_horizon: None,
gc_period: None,
image_creation_threshold: None,
pitr_interval: None,
walreceiver_connect_timeout: None,
lagging_wal_timeout: None,
max_lsn_wal_lag: None,
}
}
}
#[serde_as]
#[derive(Serialize, Deserialize, Clone)]
pub struct TenantInfo {
#[serde_as(as = "DisplayFromStr")]
pub id: TenantId,
pub state: TenantState,
pub current_physical_size: Option<u64>, // physical size is only included in `tenant_status` endpoint
pub has_in_progress_downloads: Option<bool>,
}
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct LocalTimelineInfo {
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<TimelineId>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_lsn: Option<Lsn>,
#[serde_as(as = "DisplayFromStr")]
pub last_record_lsn: Lsn,
#[serde_as(as = "Option<DisplayFromStr>")]
pub prev_record_lsn: Option<Lsn>,
#[serde_as(as = "DisplayFromStr")]
pub latest_gc_cutoff_lsn: Lsn,
#[serde_as(as = "DisplayFromStr")]
pub disk_consistent_lsn: Lsn,
pub current_logical_size: Option<u64>, // is None when timeline is Unloaded
pub current_physical_size: Option<u64>, // is None when timeline is Unloaded
pub current_logical_size_non_incremental: Option<u64>,
pub current_physical_size_non_incremental: Option<u64>,
pub wal_source_connstr: Option<String>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub last_received_msg_lsn: Option<Lsn>,
/// the timestamp (in microseconds) of the last received message
pub last_received_msg_ts: Option<u128>,
pub pg_version: u32,
}
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct RemoteTimelineInfo {
#[serde_as(as = "DisplayFromStr")]
pub remote_consistent_lsn: Lsn,
pub awaits_download: bool,
}
///
/// This represents the output of the "timeline_detail" API call.
///
#[serde_as]
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct TimelineInfo {
#[serde_as(as = "DisplayFromStr")]
pub tenant_id: TenantId,
#[serde_as(as = "DisplayFromStr")]
pub timeline_id: TimelineId,
pub local: Option<LocalTimelineInfo>,
pub remote: Option<RemoteTimelineInfo>,
}
pub type ConfigureFailpointsRequest = Vec<FailpointConfig>;
/// Information for configuring a single fail point
#[derive(Debug, Serialize, Deserialize)]
pub struct FailpointConfig {
/// Name of the fail point
pub name: String,
/// List of actions to take, using the format described in `fail::cfg`
///
/// We also support `actions = "exit"` to cause the fail point to immediately exit.
pub actions: String,
}
#[derive(Debug, Serialize, Deserialize)]
pub struct TimelineGcRequest {
pub gc_horizon: Option<u64>,
}

View File

@@ -1,7 +1,11 @@
openapi: "3.0.2"
info:
title: Page Server API
description: Neon Pageserver API
version: "1.0"
license:
name: "Apache"
url: https://github.com/neondatabase/neon/blob/main/LICENSE
servers:
- url: ""
paths:
@@ -207,6 +211,61 @@ paths:
schema:
$ref: "#/components/schemas/Error"
/v1/tenant/{tenant_id}/timeline/{timeline_id}/get_lsn_by_timestamp:
parameters:
- name: tenant_id
in: path
required: true
schema:
type: string
format: hex
- name: timeline_id
in: path
required: true
schema:
type: string
format: hex
get:
description: Get LSN by a timestamp
parameters:
- name: timestamp
in: query
required: true
schema:
type: string
format: date-time
description: A timestamp to get the LSN
responses:
"200":
description: OK
content:
application/json:
schema:
type: string
"400":
description: Error when no tenant id found in path, no timeline id or invalid timestamp
content:
application/json:
schema:
$ref: "#/components/schemas/Error"
"401":
description: Unauthorized Error
content:
application/json:
schema:
$ref: "#/components/schemas/UnauthorizedError"
"403":
description: Forbidden Error
content:
application/json:
schema:
$ref: "#/components/schemas/ForbiddenError"
"500":
description: Generic operation error
content:
application/json:
schema:
$ref: "#/components/schemas/Error"
/v1/tenant/{tenant_id}/attach:
parameters:
- name: tenant_id
@@ -556,6 +615,10 @@ components:
required:
- timeline_id
- tenant_id
- last_record_lsn
- disk_consistent_lsn
- awaits_download
- state
properties:
timeline_id:
type: string
@@ -563,33 +626,15 @@ components:
tenant_id:
type: string
format: hex
local:
$ref: "#/components/schemas/LocalTimelineInfo"
remote:
$ref: "#/components/schemas/RemoteTimelineInfo"
RemoteTimelineInfo:
type: object
required:
- awaits_download
- remote_consistent_lsn
properties:
awaits_download:
type: boolean
remote_consistent_lsn:
type: string
format: hex
LocalTimelineInfo:
type: object
required:
- last_record_lsn
- disk_consistent_lsn
properties:
last_record_lsn:
type: string
format: hex
disk_consistent_lsn:
type: string
format: hex
remote_consistent_lsn:
type: string
format: hex
ancestor_timeline_id:
type: string
format: hex
@@ -614,7 +659,41 @@ components:
format: hex
last_received_msg_ts:
type: integer
awaits_download:
type: boolean
state:
type: string
# These 'local' and 'remote' fields just duplicate some of the fields
# above. They are kept for backwards-compatibility. They can be removed,
# when the control plane has been updated to look at the above fields
# directly.
local:
$ref: "#/components/schemas/LocalTimelineInfo"
remote:
$ref: "#/components/schemas/RemoteTimelineInfo"
LocalTimelineInfo:
type: object
properties:
ancestor_timeline_id:
type: string
format: hex
ancestor_lsn:
type: string
format: hex
current_logical_size:
type: integer
current_physical_size:
type: integer
RemoteTimelineInfo:
type: object
required:
- remote_consistent_lsn
properties:
remote_consistent_lsn:
type: string
format: hex
Error:
type: object
required:

View File

@@ -12,6 +12,7 @@ use super::models::{
StatusResponse, TenantConfigRequest, TenantCreateRequest, TenantCreateResponse, TenantInfo,
TimelineCreateRequest,
};
use crate::pgdatadir_mapping::LsnForTimestamp;
use crate::storage_sync;
use crate::storage_sync::index::{RemoteIndex, RemoteTimeline};
use crate::tenant::{TenantState, Timeline};
@@ -78,13 +79,13 @@ fn get_config(request: &Request<Body>) -> &'static PageServerConf {
get_state(request).conf
}
// Helper functions to construct a LocalTimelineInfo struct for a timeline
fn local_timeline_info_from_timeline(
// Helper function to construct a TimelineInfo struct for a timeline
async fn build_timeline_info(
state: &State,
timeline: &Arc<Timeline>,
include_non_incremental_logical_size: bool,
include_non_incremental_physical_size: bool,
) -> anyhow::Result<LocalTimelineInfo> {
) -> anyhow::Result<TimelineInfo> {
let last_record_lsn = timeline.get_last_record_lsn();
let (wal_source_connstr, last_received_msg_lsn, last_received_msg_ts) = {
let guard = timeline.last_received_wal.lock().unwrap();
@@ -99,24 +100,48 @@ fn local_timeline_info_from_timeline(
}
};
let info = LocalTimelineInfo {
ancestor_timeline_id: timeline.get_ancestor_timeline_id(),
ancestor_lsn: {
match timeline.get_ancestor_lsn() {
Lsn(0) => None,
lsn @ Lsn(_) => Some(lsn),
}
},
let (remote_consistent_lsn, awaits_download) = if let Some(remote_entry) = state
.remote_index
.read()
.await
.timeline_entry(&TenantTimelineId {
tenant_id: timeline.tenant_id,
timeline_id: timeline.timeline_id,
}) {
(
Some(remote_entry.metadata.disk_consistent_lsn()),
remote_entry.awaits_download,
)
} else {
(None, false)
};
let ancestor_timeline_id = timeline.get_ancestor_timeline_id();
let ancestor_lsn = match timeline.get_ancestor_lsn() {
Lsn(0) => None,
lsn @ Lsn(_) => Some(lsn),
};
let current_logical_size = match timeline.get_current_logical_size() {
Ok(size) => Some(size),
Err(err) => {
error!("Timeline info creation failed to get current logical size: {err:?}");
None
}
};
let current_physical_size = Some(timeline.get_physical_size());
let state = timeline.current_state();
let info = TimelineInfo {
tenant_id: timeline.tenant_id,
timeline_id: timeline.timeline_id,
ancestor_timeline_id,
ancestor_lsn,
disk_consistent_lsn: timeline.get_disk_consistent_lsn(),
last_record_lsn,
prev_record_lsn: Some(timeline.get_prev_record_lsn()),
latest_gc_cutoff_lsn: *timeline.get_latest_gc_cutoff_lsn(),
current_logical_size: Some(
timeline
.get_current_logical_size()
.context("Timeline info creation failed to get current logical size")?,
),
current_physical_size: Some(timeline.get_physical_size()),
current_logical_size,
current_physical_size,
current_logical_size_non_incremental: if include_non_incremental_logical_size {
Some(timeline.get_current_logical_size_non_incremental(last_record_lsn)?)
} else {
@@ -131,32 +156,26 @@ fn local_timeline_info_from_timeline(
last_received_msg_lsn,
last_received_msg_ts,
pg_version: timeline.pg_version,
remote_consistent_lsn,
awaits_download,
state,
// Duplicate some fields in 'local' and 'remote' fields, for backwards-compatility
// with the control plane.
local: LocalTimelineInfo {
ancestor_timeline_id,
ancestor_lsn,
current_logical_size,
current_physical_size,
},
remote: RemoteTimelineInfo {
remote_consistent_lsn,
},
};
Ok(info)
}
fn list_local_timelines(
tenant_id: TenantId,
include_non_incremental_logical_size: bool,
include_non_incremental_physical_size: bool,
) -> Result<Vec<(TimelineId, LocalTimelineInfo)>> {
let tenant = tenant_mgr::get_tenant(tenant_id, true)?;
let timelines = tenant.list_timelines();
let mut local_timeline_info = Vec::with_capacity(timelines.len());
for (timeline_id, repository_timeline) in timelines {
local_timeline_info.push((
timeline_id,
local_timeline_info_from_timeline(
&repository_timeline,
include_non_incremental_logical_size,
include_non_incremental_physical_size,
)?,
))
}
Ok(local_timeline_info)
}
// healthcheck handler
async fn status_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
let config = get_config(&request);
@@ -168,6 +187,8 @@ async fn timeline_create_handler(mut request: Request<Body>) -> Result<Response<
let request_data: TimelineCreateRequest = json_request(&mut request).await?;
check_permission(&request, Some(tenant_id))?;
let state = get_state(&request);
let tenant = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
let new_timeline_info = async {
match tenant.create_timeline(
@@ -178,14 +199,10 @@ async fn timeline_create_handler(mut request: Request<Body>) -> Result<Response<
).await {
Ok(Some(new_timeline)) => {
// Created. Construct a TimelineInfo for it.
let local_info = local_timeline_info_from_timeline(&new_timeline, false, false)
let timeline_info = build_timeline_info(state, &new_timeline, false, false)
.await
.map_err(ApiError::InternalServerError)?;
Ok(Some(TimelineInfo {
tenant_id,
timeline_id: new_timeline.timeline_id,
local: Some(local_info),
remote: None,
}))
Ok(Some(timeline_info))
}
Ok(None) => Ok(None), // timeline already exists
Err(err) => Err(ApiError::InternalServerError(err)),
@@ -208,6 +225,8 @@ async fn timeline_list_handler(request: Request<Body>) -> Result<Response<Body>,
query_param_present(&request, "include-non-incremental-physical-size");
check_permission(&request, Some(tenant_id))?;
let state = get_state(&request);
let timelines = tokio::task::spawn_blocking(move || {
let _enter = info_span!("timeline_list", tenant = %tenant_id).entered();
let tenant = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
@@ -217,36 +236,18 @@ async fn timeline_list_handler(request: Request<Body>) -> Result<Response<Body>,
.map_err(|e: JoinError| ApiError::InternalServerError(e.into()))??;
let mut response_data = Vec::with_capacity(timelines.len());
for (timeline_id, timeline) in timelines {
let local = match local_timeline_info_from_timeline(
for timeline in timelines {
let timeline_info = build_timeline_info(
state,
&timeline,
include_non_incremental_logical_size,
include_non_incremental_physical_size,
) {
Ok(local) => Some(local),
Err(e) => {
error!("Failed to convert tenant timeline {timeline_id} into the local one: {e:?}");
None
}
};
)
.await
.context("Failed to convert tenant timeline {timeline_id} into the local one: {e:?}")
.map_err(ApiError::InternalServerError)?;
response_data.push(TimelineInfo {
tenant_id,
timeline_id,
local,
remote: get_state(&request)
.remote_index
.read()
.await
.timeline_entry(&TenantTimelineId {
tenant_id,
timeline_id,
})
.map(|remote_entry| RemoteTimelineInfo {
remote_consistent_lsn: remote_entry.metadata.disk_consistent_lsn(),
awaits_download: remote_entry.awaits_download,
}),
})
response_data.push(timeline_info);
}
json_response(StatusCode::OK, response_data)
@@ -265,6 +266,23 @@ fn query_param_present(request: &Request<Body>, param: &str) -> bool {
.unwrap_or(false)
}
fn get_query_param(request: &Request<Body>, param_name: &str) -> Result<String, ApiError> {
request.uri().query().map_or(
Err(ApiError::BadRequest(anyhow!("empty query in request"))),
|v| {
url::form_urlencoded::parse(v.as_bytes())
.into_owned()
.find(|(k, _)| k == param_name)
.map_or(
Err(ApiError::BadRequest(anyhow!(
"no {param_name} specified in query parameters"
))),
|(_, v)| Ok(v),
)
},
)
}
async fn timeline_detail_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&request, "tenant_id")?;
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
@@ -274,59 +292,59 @@ async fn timeline_detail_handler(request: Request<Body>) -> Result<Response<Body
query_param_present(&request, "include-non-incremental-physical-size");
check_permission(&request, Some(tenant_id))?;
let (local_timeline_info, remote_timeline_info) = async {
let state = get_state(&request);
let timeline_info = async {
let timeline = tokio::task::spawn_blocking(move || {
tenant_mgr::get_tenant(tenant_id, true)?.get_timeline(timeline_id)
tenant_mgr::get_tenant(tenant_id, true)?.get_timeline(timeline_id, false)
})
.await
.map_err(|e: JoinError| ApiError::InternalServerError(e.into()))?;
let local_timeline_info = match timeline.and_then(|timeline| {
local_timeline_info_from_timeline(
&timeline,
include_non_incremental_logical_size,
include_non_incremental_physical_size,
)
}) {
Ok(local_info) => Some(local_info),
Err(e) => {
error!("Failed to get local timeline info: {e:#}");
None
}
};
let timeline = timeline.map_err(ApiError::NotFound)?;
let remote_timeline_info = {
let remote_index_read = get_state(&request).remote_index.read().await;
remote_index_read
.timeline_entry(&TenantTimelineId {
tenant_id,
timeline_id,
})
.map(|remote_entry| RemoteTimelineInfo {
remote_consistent_lsn: remote_entry.metadata.disk_consistent_lsn(),
awaits_download: remote_entry.awaits_download,
})
};
Ok::<_, ApiError>((local_timeline_info, remote_timeline_info))
let timeline_info = build_timeline_info(
state,
&timeline,
include_non_incremental_logical_size,
include_non_incremental_physical_size,
)
.await
.context("Failed to get local timeline info: {e:#}")
.map_err(ApiError::InternalServerError)?;
Ok::<_, ApiError>(timeline_info)
}
.instrument(info_span!("timeline_detail", tenant = %tenant_id, timeline = %timeline_id))
.await?;
if local_timeline_info.is_none() && remote_timeline_info.is_none() {
Err(ApiError::NotFound(anyhow!(
"Timeline {tenant_id}/{timeline_id} is not found neither locally nor remotely"
)))
} else {
json_response(
StatusCode::OK,
TimelineInfo {
tenant_id,
timeline_id,
local: local_timeline_info,
remote: remote_timeline_info,
},
)
}
json_response(StatusCode::OK, timeline_info)
}
async fn get_lsn_by_timestamp_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&request, "tenant_id")?;
check_permission(&request, Some(tenant_id))?;
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
let timestamp_raw = get_query_param(&request, "timestamp")?;
let timestamp = humantime::parse_rfc3339(timestamp_raw.as_str())
.with_context(|| format!("Invalid time: {:?}", timestamp_raw))
.map_err(ApiError::BadRequest)?;
let timestamp_pg = postgres_ffi::to_pg_timestamp(timestamp);
let timeline = tenant_mgr::get_tenant(tenant_id, true)
.and_then(|tenant| tenant.get_timeline(timeline_id, true))
.map_err(ApiError::NotFound)?;
let result = match timeline
.find_lsn_for_timestamp(timestamp_pg)
.map_err(ApiError::InternalServerError)?
{
LsnForTimestamp::Present(lsn) => format!("{lsn}"),
LsnForTimestamp::Future(_lsn) => "future".into(),
LsnForTimestamp::Past(_lsn) => "past".into(),
LsnForTimestamp::NoData(_lsn) => "nodata".into(),
};
json_response(StatusCode::OK, result)
}
// TODO makes sense to provide tenant config right away the same way as it handled in tenant_create
@@ -337,9 +355,16 @@ async fn tenant_attach_handler(request: Request<Body>) -> Result<Response<Body>,
info!("Handling tenant attach {tenant_id}");
tokio::task::spawn_blocking(move || match tenant_mgr::get_tenant(tenant_id, false) {
Ok(_) => Err(ApiError::Conflict(
"Tenant is already present locally".to_owned(),
)),
Ok(tenant) => {
if tenant.list_timelines().is_empty() {
info!("Attaching to tenant {tenant_id} with zero timelines");
Ok(())
} else {
Err(ApiError::Conflict(
"Tenant is already present locally".to_owned(),
))
}
}
Err(_) => Ok(()),
})
.await
@@ -362,7 +387,7 @@ async fn tenant_attach_handler(request: Request<Body>) -> Result<Response<Body>,
}
return json_response(StatusCode::ACCEPTED, ());
}
// no tenant in the index, release the lock to make the potentially lengthy download opetation
// no tenant in the index, release the lock to make the potentially lengthy download operation
drop(index_accessor);
// download index parts for every tenant timeline
@@ -514,36 +539,27 @@ async fn tenant_status(request: Request<Body>) -> Result<Response<Body>, ApiErro
false
});
let tenant_state = match tenant {
Ok(tenant) => tenant.current_state(),
let (tenant_state, current_physical_size) = match tenant {
Ok(tenant) => {
let timelines = tenant.list_timelines();
// Calculate total physical size of all timelines
let mut current_physical_size = 0;
for timeline in timelines {
current_physical_size += timeline.get_physical_size();
}
(tenant.current_state(), Some(current_physical_size))
}
Err(e) => {
error!("Failed to get local tenant state: {e:#}");
if has_in_progress_downloads {
TenantState::Paused
(TenantState::Paused, None)
} else {
TenantState::Broken
(TenantState::Broken, None)
}
}
};
let current_physical_size =
match tokio::task::spawn_blocking(move || list_local_timelines(tenant_id, false, false))
.await
.map_err(|e: JoinError| ApiError::InternalServerError(e.into()))?
{
Err(err) => {
// Getting local timelines can fail when no local tenant directory is on disk (e.g, when tenant data is being downloaded).
// In that case, put a warning message into log and operate normally.
warn!("Failed to get local timelines for tenant {tenant_id}: {err}");
None
}
Ok(local_timeline_infos) => Some(
local_timeline_infos
.into_iter()
.fold(0, |acc, x| acc + x.1.current_physical_size.unwrap()),
),
};
json_response(
StatusCode::OK,
TenantInfo {
@@ -732,7 +748,7 @@ async fn tenant_config_handler(mut request: Request<Body>) -> Result<Response<Bo
json_response(StatusCode::OK, ())
}
#[cfg(any(feature = "testing", feature = "failpoints"))]
#[cfg(feature = "testing")]
async fn failpoints_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
if !fail::has_failpoints() {
return Err(ApiError::BadRequest(anyhow!(
@@ -766,11 +782,6 @@ async fn failpoints_handler(mut request: Request<Body>) -> Result<Response<Body>
}
// Run GC immediately on given timeline.
// FIXME: This is just for tests. See test_runner/regress/test_gc.py.
// This probably should require special authentication or a global flag to
// enable, I don't think we want to or need to allow regular clients to invoke
// GC.
// @hllinnaka in commits ec44f4b29, 3aca717f3
#[cfg(feature = "testing")]
async fn timeline_gc_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&request, "tenant_id")?;
@@ -778,16 +789,16 @@ async fn timeline_gc_handler(mut request: Request<Body>) -> Result<Response<Body
check_permission(&request, Some(tenant_id))?;
// FIXME: currently this will return a 500 error on bad tenant id; it should be 4XX
let repo = tenant_mgr::get_tenant(tenant_id, false).map_err(ApiError::NotFound)?;
let tenant = tenant_mgr::get_tenant(tenant_id, false).map_err(ApiError::NotFound)?;
let gc_req: TimelineGcRequest = json_request(&mut request).await?;
let _span_guard =
info_span!("manual_gc", tenant = %tenant_id, timeline = %timeline_id).entered();
let gc_horizon = gc_req.gc_horizon.unwrap_or_else(|| repo.get_gc_horizon());
let gc_horizon = gc_req.gc_horizon.unwrap_or_else(|| tenant.get_gc_horizon());
// Use tenant's pitr setting
let pitr = repo.get_pitr_interval();
let result = repo
let pitr = tenant.get_pitr_interval();
let result = tenant
.gc_iteration(Some(timeline_id), gc_horizon, pitr, true)
// FIXME: `gc_iteration` can return an error for multiple reasons; we should handle it
// better once the types support it.
@@ -796,19 +807,15 @@ async fn timeline_gc_handler(mut request: Request<Body>) -> Result<Response<Body
}
// Run compaction immediately on given timeline.
// FIXME This is just for tests. Don't expect this to be exposed to
// the users or the api.
// @dhammika in commit a0781f229
#[cfg(feature = "testing")]
async fn timeline_compact_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
let tenant_id: TenantId = parse_request_param(&request, "tenant_id")?;
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
check_permission(&request, Some(tenant_id))?;
let repo = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
let timeline = repo
.get_timeline(timeline_id)
.with_context(|| format!("No timeline {timeline_id} in repository for tenant {tenant_id}"))
let tenant = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
let timeline = tenant
.get_timeline(timeline_id, true)
.map_err(ApiError::NotFound)?;
timeline.compact().map_err(ApiError::InternalServerError)?;
@@ -822,10 +829,9 @@ async fn timeline_checkpoint_handler(request: Request<Body>) -> Result<Response<
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
check_permission(&request, Some(tenant_id))?;
let repo = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
let timeline = repo
.get_timeline(timeline_id)
.with_context(|| format!("No timeline {timeline_id} in repository for tenant {tenant_id}"))
let tenant = tenant_mgr::get_tenant(tenant_id, true).map_err(ApiError::NotFound)?;
let timeline = tenant
.get_timeline(timeline_id, true)
.map_err(ApiError::NotFound)?;
timeline
.checkpoint(CheckpointConfig::Forced)
@@ -901,6 +907,10 @@ pub fn make_router(
"/v1/tenant/:tenant_id/timeline/:timeline_id",
timeline_detail_handler,
)
.get(
"/v1/tenant/:tenant_id/timeline/:timeline_id/get_lsn_by_timestamp",
get_lsn_by_timestamp_handler,
)
.put(
"/v1/tenant/:tenant_id/timeline/:timeline_id/do_gc",
testing_api!("run timeline GC", timeline_gc_handler),

View File

@@ -12,10 +12,10 @@ use tracing::*;
use walkdir::WalkDir;
use crate::pgdatadir_mapping::*;
use crate::reltag::{RelTag, SlruKind};
use crate::tenant::Timeline;
use crate::walingest::WalIngest;
use crate::walrecord::DecodedWALRecord;
use pageserver_api::reltag::{RelTag, SlruKind};
use postgres_ffi::pg_constants;
use postgres_ffi::relfile_utils::*;
use postgres_ffi::waldecoder::WalStreamDecoder;
@@ -43,19 +43,19 @@ pub fn get_lsn_from_controlfile(path: &Path) -> Result<Lsn> {
/// The code that deals with the checkpoint would not work right if the
/// cluster was not shut down cleanly.
pub fn import_timeline_from_postgres_datadir(
path: &Path,
tline: &Timeline,
lsn: Lsn,
pgdata_path: &Path,
pgdata_lsn: Lsn,
) -> Result<()> {
let mut pg_control: Option<ControlFileData> = None;
// TODO this shoud be start_lsn, which is not necessarily equal to end_lsn (aka lsn)
// Then fishing out pg_control would be unnecessary
let mut modification = tline.begin_modification(lsn);
let mut modification = tline.begin_modification(pgdata_lsn);
modification.init_empty()?;
// Import all but pg_wal
let all_but_wal = WalkDir::new(path)
let all_but_wal = WalkDir::new(pgdata_path)
.into_iter()
.filter_entry(|entry| !entry.path().ends_with("pg_wal"));
for entry in all_but_wal {
@@ -63,7 +63,7 @@ pub fn import_timeline_from_postgres_datadir(
let metadata = entry.metadata().expect("error getting dir entry metadata");
if metadata.is_file() {
let absolute_path = entry.path();
let relative_path = absolute_path.strip_prefix(path)?;
let relative_path = absolute_path.strip_prefix(pgdata_path)?;
let file = File::open(absolute_path)?;
let len = metadata.len() as usize;
@@ -84,7 +84,7 @@ pub fn import_timeline_from_postgres_datadir(
"Postgres cluster was not shut down cleanly"
);
ensure!(
pg_control.checkPointCopy.redo == lsn.0,
pg_control.checkPointCopy.redo == pgdata_lsn.0,
"unexpected checkpoint REDO pointer"
);
@@ -92,10 +92,10 @@ pub fn import_timeline_from_postgres_datadir(
// this reads the checkpoint record itself, advancing the tip of the timeline to
// *after* the checkpoint record. And crucially, it initializes the 'prev_lsn'.
import_wal(
&path.join("pg_wal"),
&pgdata_path.join("pg_wal"),
tline,
Lsn(pg_control.checkPointCopy.redo),
lsn,
pgdata_lsn,
)?;
Ok(())

View File

@@ -5,10 +5,10 @@ pub mod import_datadir;
pub mod keyspace;
pub mod metrics;
pub mod page_cache;
pub mod page_image_cache;
pub mod page_service;
pub mod pgdatadir_mapping;
pub mod profiling;
pub mod reltag;
pub mod repository;
pub mod storage_sync;
pub mod task_mgr;
@@ -46,6 +46,8 @@ pub const DELTA_FILE_MAGIC: u16 = 0x5A61;
pub const LOG_FILE_NAME: &str = "pageserver.log";
static ZERO_PAGE: bytes::Bytes = bytes::Bytes::from_static(&[0u8; 8192]);
/// Config for the Repository checkpointer
#[derive(Debug, Clone, Copy)]
pub enum CheckpointConfig {
@@ -119,32 +121,6 @@ impl<T> TenantTimelineValues<T> {
fn new() -> Self {
Self(HashMap::new())
}
fn with_capacity(capacity: usize) -> Self {
Self(HashMap::with_capacity(capacity))
}
/// A convenience method to map certain values and omit some of them, if needed.
/// Tenants that won't have any timeline entries due to the filtering, will still be preserved
/// in the structure.
fn filter_map<F, NewT>(self, map: F) -> TenantTimelineValues<NewT>
where
F: Fn(T) -> Option<NewT>,
{
let capacity = self.0.len();
self.0.into_iter().fold(
TenantTimelineValues::<NewT>::with_capacity(capacity),
|mut new_values, (tenant_id, old_values)| {
let new_timeline_values = new_values.0.entry(tenant_id).or_default();
for (timeline_id, old_value) in old_values {
if let Some(new_value) = map(old_value) {
new_timeline_values.insert(timeline_id, new_value);
}
}
new_values
},
)
}
}
/// A suffix to be used during file sync from the remote storage,
@@ -181,35 +157,3 @@ mod backoff_defaults_tests {
);
}
}
#[cfg(test)]
mod tests {
use crate::tenant::harness::TIMELINE_ID;
use super::*;
#[test]
fn tenant_timeline_value_mapping() {
let first_tenant = TenantId::generate();
let second_tenant = TenantId::generate();
assert_ne!(first_tenant, second_tenant);
let mut initial = TenantTimelineValues::new();
initial
.0
.entry(first_tenant)
.or_default()
.insert(TIMELINE_ID, "test_value");
let _ = initial.0.entry(second_tenant).or_default();
assert_eq!(initial.0.len(), 2, "Should have entries for both tenants");
let filtered = initial.filter_map(|_| None::<&str>).0;
assert_eq!(
filtered.len(),
2,
"Should have entries for both tenants even after filtering away all entries"
);
assert!(filtered.contains_key(&first_tenant));
assert!(filtered.contains_key(&second_tenant));
}
}

Some files were not shown because too many files have changed in this diff Show More