| From: | Fujii Masao <masao(dot)fujii(at)gmail(dot)com> |
|---|---|
| To: | shihao zhong <zhong950419(at)gmail(dot)com> |
| Cc: | kehan5800(at)gmail(dot)com, pgsql-bugs(at)lists(dot)postgresql(dot)org |
| Subject: | Re: BUG #19775: Incremental backup without data checksums restores a stale visibility map page after VACUUM truncati |
| Date: | 2026-11-04 13:55:33 |
| Message-ID: | CAHGQGwGEaP7aJpkqWHFOYbPbh8QDcYKm0wOtg80HxzD_3VFgMQ@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
On Fri, Oct 23, 2026 at 12:01 PM shihao zhong <zhong950419(at)gmail(dot)com> wrote:
> Yours covers both and adds no
> WAL. So yours fits the back branches better. Whether the master should
> WAL-log the page is your call.
Agreed. I'd prefer the summarizer-side fix for that reason. I understand
the potential benefit of WAL-logging the page for other backup tools,
but I'm not yet sure how useful that would be in practice. I have no strong
objections to considering it for master later, but I'd like to apply
the summarizer-side fix first.
> 2. On 17 and 18, 013 passes even without the fix. The attached patch
> reads the VM with pg_visibility_map() per block, so it fails there
> too.
Thanks for the review and test fix! The attached v2 incorporates your
test fix and renames the test from 013_vm_truncation.pl to
014_vm_truncation.pl to avoid a numbering conflict.
Regards,
--
Fujii Masao
| Attachment | Content-Type | Size |
|---|---|---|
| v2-0001-Fix-stale-visibility-maps-in-incremental-backups-.patch | application/octet-stream | 7.1 KB |
| nocfbot-v2-0001-PG17-Fix-stale-visibility-maps-in-incremental-backups-.patch | application/octet-stream | 7.4 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | shihao zhong | 2026-11-04 14:23:11 | Re: first_value()/nth_value() with IGNORE NULLS loop forever (EXCLUDE GROUP/TIES, frame past partition end, no ORDER BY) |
| Previous Message | Peter Eisentraut | 2026-11-04 13:46:13 | Re: BUG #19830: Mention that UNIQUE on a nondeterministic ICU collation treats ignorable characters as duplicates |