Heikki Linnakangas 2e7e5f4f3a Refactor how the image layer partitioning is done ...

It got pretty ugly after the last commit. Refactor it so that we first
collect all the key ranges that need to be written out into a list of
tasks, then partition the tasks into image layers, and then write them
out. This will be much easier to parallelize, but that's not included
in this commit yet.

2024-09-13 05:41:35 +03:00

..

benches

bypass PageCache for InMemoryLayer + avoid Value::deser on L0 flush (#8537)

2024-08-28 18:31:41 +00:00

client

Implement archival_config timeline endpoint in the storage controller (#8680)

2024-09-02 13:51:45 +02:00

compaction

Fix nightly warnings 2024 june (#8151)

2024-07-12 13:58:04 +01:00

ctl

fix(pageserver): reject non-i128 key on the write path (#8648)

2024-08-29 10:07:05 -04:00

pagebench

pageserver: do vectored read on each dio-aligned section once (#8763)

2024-08-28 15:54:42 +01:00

src

Refactor how the image layer partitioning is done

2024-09-13 05:41:35 +03:00

test_data

pageserver: add layer visibility calculation (#8511)

2024-08-01 09:25:35 +00:00

Cargo.toml

proxy+pageserver: shared leaky bucket impl (#8539)

2024-08-29 11:26:52 +00:00