tests: flush wal before waiting for last record lsn (#11726)

## Problem
Compute may flush WAL on page boundaries, leaving some records partially
flushed for a long time.
It leads to `wait_for_last_flush_lsn` stuck waiting for this partial
LSN.
- Closes: https://github.com/neondatabase/cloud/issues/27876

## Summary of changes
- Flush WAL via CHECKPOINT after requesting current_wal_lsn to make sure
that the record we point to is flushed in full
- Use proper endpoint in
`test_timeline_detach_with_aux_files_with_detach_v1`
This commit is contained in:
Dmitrii Kovalkov
2025-05-08 14:00:45 +04:00
committed by GitHub
parent 40f32ea326
commit 7e55497e13
2 changed files with 9 additions and 2 deletions

View File

@@ -5477,6 +5477,13 @@ def wait_for_last_flush_lsn(
if last_flush_lsn is None:
last_flush_lsn = Lsn(endpoint.safe_psql("SELECT pg_current_wal_flush_lsn()")[0][0])
# The last_flush_lsn may not correspond to a record boundary.
# For example, if the compute flushed WAL on a page boundary,
# the remaining part of the record might not be flushed for a long time.
# This would prevent the pageserver from reaching last_flush_lsn promptly.
# To ensure the rest of the record reaches the pageserver quickly,
# we forcibly flush the WAL by using CHECKPOINT.
endpoint.safe_psql("CHECKPOINT")
results = []
for tenant_shard_id, pageserver in shards:

View File

@@ -1822,7 +1822,7 @@ def test_timeline_detach_with_aux_files_with_detach_v1(
endpoint2.safe_psql(
"SELECT pg_create_logical_replication_slot('test_slot_restore', 'pgoutput')"
)
lsn3 = wait_for_last_flush_lsn(env, endpoint, env.initial_tenant, branch_timeline_id)
lsn3 = wait_for_last_flush_lsn(env, endpoint2, env.initial_tenant, branch_timeline_id)
assert set(http.list_aux_files(env.initial_tenant, branch_timeline_id, lsn1).keys()) == set([])
assert set(http.list_aux_files(env.initial_tenant, branch_timeline_id, lsn3).keys()) == set(
["pg_replslot/test_slot_restore/state"]
@@ -1839,7 +1839,7 @@ def test_timeline_detach_with_aux_files_with_detach_v1(
assert all_reparented == set([])
# We need to ensure all safekeeper data are ingested before checking aux files: the API does not wait for LSN.
wait_for_last_flush_lsn(env, endpoint, env.initial_tenant, branch_timeline_id)
wait_for_last_flush_lsn(env, endpoint2, env.initial_tenant, branch_timeline_id)
assert set(http.list_aux_files(env.initial_tenant, env.initial_timeline, lsn2).keys()) == set(
["pg_replslot/test_slot_parent_1/state", "pg_replslot/test_slot_parent_2/state"]
), "main branch unaffected"