Reduce relcache access in WAL sender streaming logical changes
authorMichael Paquier <michael@paquier.xyz>
Wed, 5 Jan 2022 01:27:47 +0000 (10:27 +0900)
committerMichael Paquier <michael@paquier.xyz>
Wed, 5 Jan 2022 01:27:47 +0000 (10:27 +0900)
commit5ddfebded41f7754b3c4fc1b576ca33736efcab6
tree7f6e793fa56d1c0245a655491c7474cad0aa1602
parentf9db153c282de599a7d5ae2694f50fe10fba8461
Reduce relcache access in WAL sender streaming logical changes

get_rel_sync_entry(), which is called each time a change needs to be
logically replicated, is a rather hot code path in the WAL sender
sending logical changes.  This code path was doing a relcache access on
relkind and relpartition for each logical change, but we only need to
know this information when building or re-building the cached
information for a relation.

Some measurements prove that this is noticeable in perf profiles,
particularly when attempting to replicate changes from relations that
are not published as these cause less overhead in the WAL sender,
delaying further the replication of changes for relations that are
published.

Issue introduced in 83fd453.

Author: Hou Zhijie
Reviewed-by: Kyotaro Horiguchi, Euler Taveira
Discussion: https://postgr.es/m/OS0PR01MB5716E863AA9E591C1F010F7A947D9@OS0PR01MB5716.jpnprd01.prod.outlook.com
Backpatch-through: 13
src/backend/replication/pgoutput/pgoutput.c