public inbox for [email protected]help / color / mirror / Atom feed
Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops 5+ messages / 3 participants [nested] [flat]
* Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops @ 2026-04-10 08:38 Lakshmi N <[email protected]> 2026-04-10 11:57 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> 0 siblings, 1 reply; 5+ messages in thread From: Lakshmi N @ 2026-04-10 08:38 UTC (permalink / raw) To: [email protected] Hi hackers, I noticed CFI is missing while scanning pg_class for RELKIND_RELATION and RELKIND_TOASTVALUE. On a database with several thousands of tables, these scans can take a noticeable amount of time. Attached a patch to address this. Regards, Lakshmi Attachments: [application/octet-stream] 0001-autovacuum-cfi.patch (639B, 3-0001-autovacuum-cfi.patch) download | inline diff: diff --git a/src/backend/postmaster/autovacuum.c b/src/backend/postmaster/autovacuum.c index 82061247..90575a52 100644 --- a/src/backend/postmaster/autovacuum.c +++ b/src/backend/postmaster/autovacuum.c @@ -2042,6 +2042,8 @@ do_autovacuum(void) bool wraparound; AutoVacuumScores scores; + CHECK_FOR_INTERRUPTS(); + if (classForm->relkind != RELKIND_RELATION && classForm->relkind != RELKIND_MATVIEW) continue; @@ -2146,6 +2148,8 @@ do_autovacuum(void) bool wraparound; AutoVacuumScores scores; + CHECK_FOR_INTERRUPTS(); + /* * We cannot safely process other backends' temp tables, so skip 'em. */ ^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops 2026-04-10 08:38 Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> @ 2026-04-10 11:57 ` Dapeng Wang <[email protected]> 2026-04-11 05:02 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 0 siblings, 1 reply; 5+ messages in thread From: Dapeng Wang @ 2026-04-10 11:57 UTC (permalink / raw) To: Lakshmi N <[email protected]>; +Cc: [email protected] Lakshmi N <[email protected]> 于2026年4月10日周五 16:39写道： > Hi hackers, > > I noticed CFI is missing while scanning pg_class for RELKIND_RELATION and > RELKIND_TOASTVALUE. On a database with several thousands of tables, these > scans can take a noticeable amount of time. Attached a patch to address > this. > > Regards, > Lakshmi > Hi Lakshmi, The patch applies cleanly to HEAD and compiles without warnings. make check passes (one unrelated ICU collation diff). The two CHECK_FOR_INTERRUPTS() additions in do_autovacuum() look correct and are well-placed at the top of each loop iteration, before any resources are acquired. I noticed there are other similar catalog scan loops in autovacuum.c that also lack CHECK_FOR_INTERRUPTS(): - The pg_database scan loop around line 1854 - The pg_class scan loop around line 3664 Should those be covered as well? Regards, Dapeng Wang ^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops 2026-04-10 08:38 Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 2026-04-10 11:57 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> @ 2026-04-11 05:02 ` Lakshmi N <[email protected]> 2026-04-11 12:19 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> 0 siblings, 1 reply; 5+ messages in thread From: Lakshmi N @ 2026-04-11 05:02 UTC (permalink / raw) To: Dapeng Wang <[email protected]>; +Cc: [email protected] Hi, On Fri, Apr 10, 2026 at 4:57 AM Dapeng Wang <[email protected]> wrote: > Lakshmi N <[email protected]> 于2026年4月10日周五 16:39写道： > >> Hi hackers, >> >> I noticed CFI is missing while scanning pg_class for RELKIND_RELATION and >> RELKIND_TOASTVALUE. On a database with several thousands of tables, these >> scans can take a noticeable amount of time. Attached a patch to address >> this. >> >> Regards, >> Lakshmi >> > Hi Lakshmi, > > The patch applies cleanly to HEAD and compiles without warnings. > make check passes (one unrelated ICU collation diff). > > The two CHECK_FOR_INTERRUPTS() additions in do_autovacuum() look > correct and are well-placed at the top of each loop iteration, > before any resources are acquired. > > I noticed there are other similar catalog scan loops in > autovacuum.c that also lack CHECK_FOR_INTERRUPTS(): > > - The pg_database scan loop around line 1854 > - The pg_class scan loop around line 3664 > > Should those be covered as well? > Updated the patch covering those two as well. Regards, Lakshmi > Attachments: [application/octet-stream] v2-0001-Add-CHECK_FOR_INTERRUPTS-in-autovacuum-table-scannin.patch (1.4K, 3-v2-0001-Add-CHECK_FOR_INTERRUPTS-in-autovacuum-table-scannin.patch) download | inline diff: From d2d31987d2e479a8d92494644c183c8bd2f749b2 Mon Sep 17 00:00:00 2001 From: Lakshmi N <[email protected]> Date: Fri, 10 Apr 2026 01:39:04 -0700 Subject: [PATCH] Add CHECK_FOR_INTERRUPTS in autovacuum table-scanning loops --- src/backend/postmaster/autovacuum.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/src/backend/postmaster/autovacuum.c b/src/backend/postmaster/autovacuum.c index 82061247988..5e36137925a 100644 --- a/src/backend/postmaster/autovacuum.c +++ b/src/backend/postmaster/autovacuum.c @@ -1859,6 +1859,8 @@ get_database_list(void) avw_dbase *avdb; MemoryContext oldcxt; + CHECK_FOR_INTERRUPTS(); + /* * If database has partially been dropped, we can't, nor need to, * vacuum it. @@ -2042,6 +2044,8 @@ do_autovacuum(void) bool wraparound; AutoVacuumScores scores; + CHECK_FOR_INTERRUPTS(); + if (classForm->relkind != RELKIND_RELATION && classForm->relkind != RELKIND_MATVIEW) continue; @@ -2146,6 +2150,8 @@ do_autovacuum(void) bool wraparound; AutoVacuumScores scores; + CHECK_FOR_INTERRUPTS(); + /* * We cannot safely process other backends' temp tables, so skip 'em. */ @@ -3669,6 +3675,8 @@ pg_stat_get_autovacuum_scores(PG_FUNCTION_ARGS) Datum vals[10]; bool nulls[10] = {false}; + CHECK_FOR_INTERRUPTS(); + /* skip ineligible entries */ if (form->relkind != RELKIND_RELATION && form->relkind != RELKIND_MATVIEW && -- 2.43.0 ^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops 2026-04-10 08:38 Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 2026-04-10 11:57 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> 2026-04-11 05:02 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> @ 2026-04-11 12:19 ` Dapeng Wang <[email protected]> 2026-04-11 20:10 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Sami Imseih <[email protected]> 0 siblings, 1 reply; 5+ messages in thread From: Dapeng Wang @ 2026-04-11 12:19 UTC (permalink / raw) To: Lakshmi N <[email protected]>; +Cc: [email protected] Lakshmi N <[email protected]> 于2026年4月11日周六 13:02写道： > Hi, > > On Fri, Apr 10, 2026 at 4:57 AM Dapeng Wang <[email protected]> > wrote: > >> Lakshmi N <[email protected]> 于2026年4月10日周五 16:39写道： >> >>> Hi hackers, >>> >>> I noticed CFI is missing while scanning pg_class for RELKIND_RELATION >>> and RELKIND_TOASTVALUE. On a database with several thousands of tables, >>> these scans can take a noticeable amount of time. Attached a patch to >>> address this. >>> >>> Regards, >>> Lakshmi >>> >> Hi Lakshmi, >> >> The patch applies cleanly to HEAD and compiles without warnings. >> make check passes (one unrelated ICU collation diff). >> >> The two CHECK_FOR_INTERRUPTS() additions in do_autovacuum() look >> correct and are well-placed at the top of each loop iteration, >> before any resources are acquired. >> >> I noticed there are other similar catalog scan loops in >> autovacuum.c that also lack CHECK_FOR_INTERRUPTS(): >> >> - The pg_database scan loop around line 1854 >> - The pg_class scan loop around line 3664 >> >> Should those be covered as well? >> > > Updated the patch covering those two as well. > > > Regards, > Lakshmi > Hi Lakshmi, The updated patch now covers all four catalog scan loops. Applies cleanly, compiles without warnings, and all 247 regression tests pass. Regards, Dapeng Wang ^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops 2026-04-10 08:38 Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 2026-04-10 11:57 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> 2026-04-11 05:02 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 2026-04-11 12:19 ` Re: Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Dapeng Wang <[email protected]> @ 2026-04-11 20:10 ` Sami Imseih <[email protected]> 0 siblings, 0 replies; 5+ messages in thread From: Sami Imseih @ 2026-04-11 20:10 UTC (permalink / raw) To: Dapeng Wang <[email protected]>; +Cc: Lakshmi N <[email protected]>; [email protected] Hi, It was discussed [1] here if C_F_I is needed in these do_autovacuum/autovacuum_scores code paths, and it was decided not to add them. Also, at least on my tests with 100k evern with 1 million tables, I did not see a reason to do so. These are catalog scans that should happen quick enough, even in extreme cases, to not make this worthwhile. Others may have another opinion. -- Sami [1] [https://www.postgresql.org/message-id/acwTxpz3Toxt0ty8%40nathan] ^ permalink raw reply [nested|flat] 5+ messages in thread
end of thread, other threads:[~2026-04-11 20:10 UTC | newest] Thread overview: 5+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2026-04-10 08:38 Add missing CHECK_FOR_INTERRUPTS in autovacuum catalog scan loops Lakshmi N <[email protected]> 2026-04-10 11:57 ` Dapeng Wang <[email protected]> 2026-04-11 05:02 ` Lakshmi N <[email protected]> 2026-04-11 12:19 ` Dapeng Wang <[email protected]> 2026-04-11 20:10 ` Sami Imseih <[email protected]>
This inbox is served by agora; see mirroring instructions for how to clone and mirror all data and code used for this inbox