| From: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)postgresql(dot)org> |
| Cc: | pgsql-committers(at)postgresql(dot)org |
| Subject: | Re: pgsql: Fix a violation of WAL coding rules in the recent patch to |
| Date: | 2009-08-24 05:39:45 |
| Message-ID: | 4A9227A1.3020603@enterprisedb.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-committers |
Tom Lane wrote:
> Log Message:
> -----------
> Fix a violation of WAL coding rules in the recent patch to include an
> "all tuples visible" flag in heap page headers. The flag update *must*
> be applied before calling XLogInsert, but heap_update and the tuple
> moving routines in VACUUM FULL were ignoring this rule. A crash and
> replay could therefore leave the flag incorrectly set, causing rows
> to appear visible in seqscans when they should not be. This might explain
> recent reports of data corruption from Jeff Ross and others.
Ouch! Yeah, the old code would write full-page images with the flag
incorrectly set.
Thanks!
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Heikki Linnakangas | 2009-08-24 06:12:57 | Re: pgsql: Fix a violation of WAL coding rules in the recent patch to |
| Previous Message | Tom Lane | 2009-08-24 03:10:16 | pgsql: Run the "tablespace" regression test first not last. |