| From: | Greg Stark <stark(at)mit(dot)edu> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Noah Misch <noah(at)leadboat(dot)com>, Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Temporary tables versus wraparound... again |
| Date: | 2022-03-31 20:05:13 |
| Message-ID: | CAM-w4HP-Un_YQ0EXKzV_hqX==XXY2GwycvzsOuQZPQPs9ZjJXA@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
I've updated the patches.
Adding the assertion actually turned up a corner case where RecentXmin
was *not* set. If you lock a temporary table and that's the only thing
you do in a transaction then the flag is set indicating you've used
the temp schema but you never take a snapshot :(
I also split out the warnings and added a test that relfrozenxid was
advanced on the toast table as well.
I haven't wrapped my head around multixacts yet. It's complicated by
this same codepath being used for truncates of regular tables that
were created in the same transaction.
| Attachment | Content-Type | Size |
|---|---|---|
| v5-0002-Update-relfrozenxmin-when-truncating-temp-tables.patch | text/x-patch | 3.8 KB |
| v5-0003-Add-test-for-truncating-temp-tables-advancing-rel.patch | text/x-patch | 6.1 KB |
| v5-0001-Add-warnings-when-old-temporary-tables-are-found-.patch | text/x-patch | 9.0 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tomas Vondra | 2022-03-31 20:05:30 | Re: unlogged sequences |
| Previous Message | Larry Rosenman | 2022-03-31 20:03:50 | Re: head fails to build on SLES 12 |