From: | Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> |
---|---|
To: | Craig Ringer <craig(at)2ndquadrant(dot)com> |
Cc: | PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>, Andres Freund <andres(at)anarazel(dot)de> |
Subject: | Re: BUGFIX: standby disconnect can corrupt serialized reorder buffers |
Date: | 2018-01-05 19:28:29 |
Message-ID: | 20180105192829.wiwqjv2nmrryoco3@alvherre.pgsql |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I think this should use ReadDirExtended with an elevel less than ERROR,
and do nothing.
Why have strcmp(.) and strcmp(..)? These are going to be skipped by the
comparison to "xid" prefix anyway. Looks like strcmp processing power waste.
Please don't use bare sprintf() -- upgrade to snprintf.
With this coding, if I put a root-owned file "xidfoo" in a replslot
directory, it will PANIC the server. Is that okay? Why not read the
file name with sscanf(), since we know the precise format it has? Then
we don't need to bother with random crap left around. Maybe a good time
to put the "xid-%u-lsn-%X-%X.snap" literal in a macro? I guess the
rationale is that if you let random people put "xidfoo" files in your
replication slot dirs, you deserve a PANIC anyway, but I'm not sure.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
From | Date | Subject | |
---|---|---|---|
Next Message | Alexander Korotkov | 2018-01-05 19:31:36 | Re: [HACKERS] Pluggable storage |
Previous Message | Bruce Momjian | 2018-01-05 19:24:54 | Re: pgsql: pg_upgrade: simplify code layout in a few places |