rust/neon
1
0
Fork 0
mirror of https://github.com/neondatabase/neon.git synced 2025-12-22 21:59:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix pgcopydb seg fault and -c idle_in_transaction_session_timeout=0 (#10692)

Browse Source 
## Problem

During ingest_benchmark which uses `pgcopydb`
([see](https://github.com/dimitri/pgcopydb))we sometimes had outages.
- when PostgreSQL COPY step failed we got a segfault (reported
[here](https://github.com/dimitri/pgcopydb/issues/899))
- the root cause was Neon idle_in_transaction_session_timeout is set to
5 minutes which is suboptimal for long-running tasks like project import
(reported [here](https://github.com/dimitri/pgcopydb/issues/900))

## Summary of changes

Patch pgcopydb to avoid segfault.
override idle_in_transaction_session_timeout and set it to "unlimited"
This commit is contained in:
Peter Bendel
2025-02-06 15:39:45 +01:00
committed by GitHub
parent f22d41eaec
commit 839f41f5bb
4 changed files with 42 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
.dockerignore
View File

@@ -24,3 +24,4 @@
!storage_controller/
!vendor/postgres-*/
!workspace_hack/
!build_tools/patches

3
build-tools.Dockerfile
View File

@@ -12,6 +12,8 @@ RUN echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries && \
echo -e "retry_connrefused = on\ntimeout=15\ntries=5\n" > /root/.wgetrc && \
echo -e "--retry-connrefused\n--connect-timeout 15\n--retry 5\n--max-time 300\n" > /root/.curlrc
COPY build_tools/patches/pgcopydbv017.patch /pgcopydbv017.patch
RUN if [ "${DEBIAN_VERSION}" = "bookworm" ]; then \
set -e && \
apt update && \
@@ -44,6 +46,7 @@ RUN if [ "${DEBIAN_VERSION}" = "bookworm" ]; then \
mkdir /tmp/pgcopydb && \
tar -xzf /tmp/pgcopydb.tar.gz -C /tmp/pgcopydb --strip-components=1 && \
cd /tmp/pgcopydb && \
patch -p1 < /pgcopydbv017.patch && \
make -s clean && \
make -s -j12 install && \
libpq_path=$(find /lib /usr/lib -name "libpq.so.5" | head -n 1) && \

37
build_tools/patches/pgcopydbv017.patch Normal file
View File

@@ -0,0 +1,37 @@
diff --git a/src/bin/pgcopydb/copydb.c b/src/bin/pgcopydb/copydb.c
index d730b03..69a9be9 100644
--- a/src/bin/pgcopydb/copydb.c
+++ b/src/bin/pgcopydb/copydb.c
@@ -44,6 +44,7 @@ GUC dstSettings[] = {
{ "synchronous_commit", "'off'" },
{ "statement_timeout", "0" },
{ "lock_timeout", "0" },
+ { "idle_in_transaction_session_timeout", "0" },
{ NULL, NULL },
};
diff --git a/src/bin/pgcopydb/pgsql.c b/src/bin/pgcopydb/pgsql.c
index 94f2f46..86b9448 100644
--- a/src/bin/pgcopydb/pgsql.c
+++ b/src/bin/pgcopydb/pgsql.c
@@ -3174,11 +3174,18 @@ pgcopy_log_error(PGSQL *pgsql, PGresult *res, const char *context)
/* errors have already been logged */
return;
}
-
if (res != NULL)
{
char *sqlstate = PQresultErrorField(res, PG_DIAG_SQLSTATE);
- strlcpy(pgsql->sqlstate, sqlstate, sizeof(pgsql->sqlstate));
+ if (sqlstate == NULL)
+ {
+ // PQresultErrorField returned NULL!
+ pgsql->sqlstate[0] = '\0'; // Set to an empty string to avoid segfault
+ }
+ else
+ {
+ strlcpy(pgsql->sqlstate, sqlstate, sizeof(pgsql->sqlstate));
+ }
}
char *endpoint =

2
test_runner/performance/test_perf_ingest_using_pgcopydb.py
View File

@@ -136,7 +136,7 @@ def run_command_and_log_output(command, log_file_path: Path):
"LD_LIBRARY_PATH": f"{os.getenv('PGCOPYDB_LIB_PATH')}:{os.getenv('PG_16_LIB_PATH')}",
"PGCOPYDB_SOURCE_PGURI": cast(str, os.getenv("BENCHMARK_INGEST_SOURCE_CONNSTR")),
"PGCOPYDB_TARGET_PGURI": cast(str, os.getenv("BENCHMARK_INGEST_TARGET_CONNSTR")),
"PGOPTIONS": "-c maintenance_work_mem=8388608 -c max_parallel_maintenance_workers=7",
"PGOPTIONS": "-c idle_in_transaction_session_timeout=0 -c maintenance_work_mem=8388608 -c max_parallel_maintenance_workers=7",
}
# Combine the current environment with custom variables
env = os.environ.copy()