| From: | Álvaro Rodríguez <alvaro(at)datadoghq(dot)com> |
|---|---|
| To: | Sami Imseih <samimseih(dot)pg(at)gmail(dot)com> |
| Cc: | Zsolt Parragi <zsolt(dot)parragi(at)percona(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org, Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com>, "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com> |
| Subject: | Re: tablecmds: fix bug where index rebuild loses replica identity on partitions |
| Date: | 2026-10-06 11:37:09 |
| Message-ID: | CA+C_kKWVb5PrrGJM5Bd4rPsf-V=h20xoeCHMtPy3WCYy4=YwUQ@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi all,
I have been working on another patch [0] and was pointed to this
thread. My changes are focused on avoiding preventable rebuilds of the
leaf indexes on partitioned tables, when we do ALTER TABLE ALTER
COLUMN TYPE. I change the order in which indexes are processed and do
some other minor changes to ensure that they are correctly remembered
on rebuild.
I believe commit 0003 here addresses a subset of the issues that my
patch fixes. To validate, I have combined 0001 + 0002 + my patch +
0003's tests and the tests seem to be passing (the only change in test
output is the at_partitioned test, where I get the storage changes
that my patch is fixing). Commits 0001 and 0002 are not addressed by
my patch at all, as far as I can tell. I wonder if it's worth joining
efforts!
Best,
Álvaro
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Matthias van de Meent | 2026-10-06 11:48:57 | Re: Exploring pass-by-value for small Bitmapset sets |
| Previous Message | Zhijie Hou | 2026-10-06 11:27:27 | Re: Parallel Apply |