Re: SLRU statistics

From: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
To: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
Cc: Masahiko Sawada <masahiko(dot)sawada(at)2ndquadrant(dot)com>, "tsunakawa(dot)takay(at)fujitsu(dot)com" <tsunakawa(dot)takay(at)fujitsu(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: SLRU statistics
Date: 2020-03-27 00:41:11
Message-ID: 20200327004111.md4suoi6u5suv2qe@development
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

Attached is v3 of the patch with one big change and various small ones.

The main change is that it gets rid of the SlruType enum and the new
field in SlruCtlData. Instead, the patch now uses the name passed to
SimpleLruInit (which is then stored as LWLock tranche name).

The counters are still stored in a fixed-sized array, and there's a
simple name/index mapping. We don't have any registry of stable SLRU
IDs, so I can't think of anything better, and I think this is good
enough for now.

The other change is that I got rid of the io_error counter. We don't
have that for shared buffers etc. either, anyway.

I've also renamed the colunms from "pages" to "blks" to make it
consistent with other similar stats (blks_hit, blks_read). I've
renamed the fields to "blks_written" and "blks_zeroed".

And finally, I've added the view to monitoring.sgml.

Barring objections, I'll get this committed in the next few days, after
reviewing the comments a bit.

regards

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

Attachment Content-Type Size
0001-SLRU-stats-v3.patch text/plain 25.5 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Alvaro Herrera 2020-03-27 00:41:21 Re: error context for vacuum to include block number
Previous Message Alvaro Herrera 2020-03-27 00:39:17 Re: allow online change primary_conninfo