| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
| Cc: | Andrew Dunstan <andrew(at)dunslane(dot)net>, Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Status of autovacuum and the sporadic stats failures ? |
| Date: | 2007-02-07 23:13:39 |
| Message-ID: | 12305.1170890019@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> FWIW I'm thinking that the corresponding code for handling the backends'
> stats could be simplified, removing the hack that stores it in
> TopTransactionContext, and just having a call to the stats flush
> function in AbortTransaction and CommitTransaction.
Yeah, thanks for the idea. The pgstats code itself is now decoupled
from transaction boundaries. It still has one hack to know that
autovacuum wants to keep its stats for the whole run, but it'll probably
be a lot easier to whack it around the way you want now.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | FAST PostgreSQL | 2007-02-07 23:36:44 | Re: [pgsql-patches] pg_get_domaindef |
| Previous Message | Merlin Moncure | 2007-02-07 21:52:29 | Re: HOT for PostgreSQL 8.3 |