From: | Simon Riggs <simon(at)2ndQuadrant(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: A note about hash-based catcache invalidations |
Date: | 2011-08-17 07:42:11 |
Message-ID: | CA+U5nML8ORSL7Vg7A3Zb=N8ywb1Y2nU1+-4yVatZXgu9d2CRug@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Tue, Aug 16, 2011 at 10:17 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Any objections to that plan?
None at all, but some questions.
This overhaul of the cache mechanism has been extensive, so you're now
very well placed to answer related questions.
As you know, I've been trying to reduce the lock strength of some DDL
operations. When that was last discussed there were two "options". The
first was to re-write SnapshotNow, which in my opinion is necessary
but solves only part of the problem. I proposed explicit locking
around catalog access, which would affect the cache path/code. I don't
like that, but I don't see another way.
From where you are now, do have any insight about how to tackle the
locking problem? Thanks.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2011-08-17 08:49:21 | Re: Enforcing that all WAL has been replayed after restoring from backup |
Previous Message | Alexander Korotkov | 2011-08-17 07:11:34 | Re: WIP: Fast GiST index build |