Horton

Re: Delay locking partitions during INSERT and UPDATE

Tomas Vondra <tomas.vondra@2ndquadrant.com>

View thread View on postgr.es
From: Tomas Vondra <tomas.vondra@2ndquadrant.com>
To: John Naylor <john.naylor@2ndquadrant.com>, David Rowley <david.rowley@2ndquadrant.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2019-01-24T21:17:12Z
Lists: pgsql-hackers 
On 1/24/19 9:50 PM, John Naylor wrote:
> On Wed, Jan 23, 2019 at 7:56 PM David Rowley
> <david.rowley@2ndquadrant.com> wrote:
>> On Thu, 24 Jan 2019 at 13:38, John Naylor <john.naylor@2ndquadrant.com> wrote:
>>> Hmm, now instead of an 85x speedup over master in the 10k partition
>>> case, I only get 20x. Anyone else see this?
>>
>> What's it like with fsync off?
> 
> No change. Just in case, I tried git bisect and also recreated the
> cluster, but never got the same performance as my first test, so not
> sure what happened.

I can still see about the same performance as before (on both clusters).

regards

-- 
Tomas Vondra                  http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Commits

  1. Delay lock acquisition for partitions until we route a tuple to them.