| From: | Mark Dilger <mark(dot)dilger(at)enterprisedb(dot)com> |
|---|---|
| To: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: data corruption hazard in reorderbuffer.c |
| Date: | 2021-07-15 20:51:41 |
| Message-ID: | 4F3BC7FF-4748-4A12-8B0F-B64B8F658919@enterprisedb.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> On Jul 15, 2021, at 1:03 PM, Mark Dilger <mark(dot)dilger(at)enterprisedb(dot)com> wrote:
>
> Skipping some writes while not others easily creates a variety of failures, and for brevity I won't post a patch to demonstrate that here.
If anybody is curious, one common error I see when simulating a close() skipping partial changes rather than whole ones looks like:
ERROR: got sequence entry 31 for toast chunk 16719 instead of seq 21
(where the exact numbers differ, of course). This symptom has shown up in at least two ([1], [2] below) unsolved user bug reports specifically mentioning replication. That doesn't prove a connection between the those reports and this issue, but it makes me wonder.
[1] https://www.postgresql.org/message-id/CA+_m4OBs2aPkjqd1gxnx2ykuTJogYCfq8TZgr1uPP3ZtBTyvew@mail.gmail.com
[2] https://www.postgresql.org/message-id/flat/3bd6953d-3da6-040c-62bf-9522808d5c2f%402ndquadrant.com#f6f165ebea024f851d47a17723de5d29
—
Mark Dilger
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Mark Dilger | 2021-07-15 21:17:32 | Re: data corruption hazard in reorderbuffer.c |
| Previous Message | David G. Johnston | 2021-07-15 20:24:17 | Re: pg_upgrade does not upgrade pg_stat_statements properly |