From: | Justin Pryzby <pryzby(at)telsasoft(dot)com> |
---|---|
To: | Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> |
Cc: | Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>, Julien Rouhaud <rjuju123(at)gmail(dot)com>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: backend type in log_line_prefix? |
Date: | 2020-03-15 09:57:28 |
Message-ID: | 20200315095728.GA26184@telsasoft.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Fri, Mar 13, 2020 at 10:22:52PM +0100, Peter Eisentraut wrote:
> >Can I suggest:
> >
> >- appendCSVLiteral(&buf, MyBgworkerEntry->bgw_type);
> >+ appendCSVLiteral(&buf, MyBgworkerEntry->bgw_name);
>
> The difference is intentional. bgw_type is so that you can filter and group
> by type. The bgw_name could be totally different for each instance.
I found 5373bc2a0867048bb78f93aede54ac1309b5e227
Your patch adds bgw_type, which is also in pg_stat_activity, so I agree it's
good to allow include it log_line_prefix and CSV.
I suggest the CSV/log should also have the leader_pid, corresponding to
| b025f32e0b Add leader_pid to pg_stat_activity
With the attached on top of your patch, CSV logs like:
2020-03-14 22:09:39.395 CDT,"pryzbyj","template1",17030,"[local]",5e6d9c69.4286,2,"idle",2020-03-14 22:09:29 CDT,3/23,0,LOG,00000,"statement: explain analyze SELECT COUNT(1), a.a FROM t a JOIN t b ON a.a=b.a GROUP BY 2;",,,,,,,,,"psql","client backend",
2020-03-14 22:09:43.094 CDT,,,17042,,5e6d9c73.4292,1,,2020-03-14 22:09:39 CDT,4/3,0,LOG,00000,"temporary file: path ""base/pgsql_tmp/pgsql_tmp17042.0"", size 4694016",,,,,,"explain analyze SELECT COUNT(1), a.a FROM t a JOIN t b ON a.a=b.a GROUP BY 2;",,,"psql","parallel worker",17030
2020-03-14 22:09:43.094 CDT,,,17043,,5e6d9c73.4293,1,,2020-03-14 22:09:39 CDT,5/3,0,LOG,00000,"temporary file: path ""base/pgsql_tmp/pgsql_tmp17043.0"", size 4694016",,,,,,"explain analyze SELECT COUNT(1), a.a FROM t a JOIN t b ON a.a=b.a GROUP BY 2;",,,"psql","parallel worker",17030
As for my question "what's using/trying/failing to use parallel workers", I was
able to look into that by parsing "Workers Planned/Launched" from autoexplain.
It's not a *good* way to do it, but I don't see how to do better and I don't
see any way this patch can improve that.
--
Justin
Attachment | Content-Type | Size |
---|---|---|
v1-0001-Fix-previous-commit.patch | text/x-diff | 1.9 KB |
v1-0002-Include-the-leader-PID-in-logfile.patch | text/x-diff | 5.8 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Justin Pryzby | 2020-03-15 10:01:50 | Re: Berserk Autovacuum (let's save next Mandrill) |
Previous Message | Dilip Kumar | 2020-03-15 07:45:19 | Re: [HACKERS] Moving relation extension locks out of heavyweight lock manager |