Re: improve DEBUG1 logging of parallel workers for CREATE INDEX?

From: Guillaume Lelarge <guillaume(at)lelarge(dot)info>
To: Sami Imseih <samimseih(at)gmail(dot)com>
Cc: Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: improve DEBUG1 logging of parallel workers for CREATE INDEX?
Date: 2025-01-08 20:21:46
Message-ID: CAECtzeVgxsxHSXhbB=UQJXN5ifUTEEQTJZXB00NQj2X25e=-jg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

Le mer. 8 janv. 2025 à 18:37, Sami Imseih <samimseih(at)gmail(dot)com> a écrit :

> > s/parallel vacuum workers for index vacuuming/parallel workers for index
> creation/?
>
> oops, that's my oversight from copying the message from vacuum. fixed.
>
> > I'd add "in parallel" to match its counterpart "serially" above. That
> would
> > make it more clear in case one just look for "building index" in the log.
>
> good point.
>
> Below is the new output with the attached v2.
>
> postgres=# create index foo_idx1 on foo(id);
> DEBUG: building index "foo_idx1" on table "foo" with parallel workers
> DEBUG: launched 1 parallel workers for index creation (planned: 1)
> DEBUG: index "foo_idx1" can safely use deduplication
> CREATE INDEX
>
>
I tend to agree it might be better than Benoit's patch on the index
messages, though I'm afraid that DEBUG1 level won't work for many users.
DEBUGx are for developers, not users. A better log level for me would be
LOG. For client messages, LOG messages won't be displayed by default. So
there's still a need for a "SET client_min_messages to LOG" but it won't be
necessary to go all the way to DEBUG1.

Regards.

--
Guillaume.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2025-01-08 20:23:02 Re: Fwd: Re: A new look at old NFS readdir() problems?
Previous Message Tom Lane 2025-01-08 20:21:27 Re: Fwd: Re: A new look at old NFS readdir() problems?