[ClusterLabs] Antw: [EXT] Feedback wanted: using the systemd message catalog

[Ulrich Windl]

Hi Ken,

personally I think systemd is already logging too much, and I don't think that
adding instructions to many log messages is actually helpful (It could be done
as separate log message (maybe at severity info) already).
In Windows I see the problem that it's very hard to find real problems as the
logs are flooded with all kinds of stuff.

Example (being rather useless):
Feb 17 12:01:01 h03 systemd[1]: Started Session 41667 of user root.
-- Subject: Unit session-41667.scope has finished start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel 
--
-- Unit session-41667.scope has finished starting up.
--
-- The start-up result is done.

So is there any useful info in those extra lines?

Regards,
Ulrich



>>> Ken Gaillot <kgaillot at redhat.com> schrieb am 16.02.2021 um 20:47 in
Nachricht
<8a2c91037be49816055a717583c6c505c996f5e7.camel at redhat.com>:
> Hi all,
> 
> The systemd journal has a feature called the message catalog, which
> allows "journalctl ‑x" to show extended troubleshooting hints along
> with log messages. Apparently a similar feature is familiar to many
> Windows administrators.
> 
> So far, I've only found systemd itself that uses the feature. As an
> example, if the usual log message is a one‑liner about httpd dumping
> core, with journalctl ‑x it would look like:
> 
>  ‑‑ fc2e22bc6ee647b6b90729ab34a250b1
>  Subject: Process 1234 (httpd) dumped core
>  Defined‑By: systemd
>  Support: http://lists.freedesktop.org/mailman/listinfo/systemd‑devel 
>  Documentation: man:core(5)
>  Documentation: 
>
http://www.freedesktop.org/wiki/Software/systemd/catalog/fc2e22bc6ee647b6b907

> 29ab34a250b1
> 
>  Process 1234 (httpd) crashed and dumped core.
> 
>  This usually indicates a programming error in the crashing program and
>  should be reported to its vendor as a bug.
> 
> There is currently a proposed pull request ( 
> https://github.com/ClusterLabs/libqb/pull/433 ) to add support for this
> to libqb's logging API, which the rest of the cluster software could
> use.
> 
> As an example, if someone forcibly uploads a CIB with syntax errors,
> Pacemaker will currently log "The cluster is mis‑configured ‑ shutting
> down and staying down" and exit. We could make it so that the
> journalctl ‑x entry tells the user how to use crm_verify to figure out
> what's wrong, how to use cibadmin to correct any errors, and a link to
> Pacemaker Explained as a configuration reference.
> 
> This would involve considerable work for developers, so I'm curious how
> many users would find this useful and would use it. Especially if most
> of the time you used journalctl ‑x, there was no extended information,
> but occasionally there was.
> ‑‑ 
> Ken Gaillot <kgaillot at redhat.com>
> 
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users 
> 
> ClusterLabs home: https://www.clusterlabs.org/