Credentials not working after restart of Server

NethServer release 7.3.1611
SSSD
Did anybody else had the expirience, that after a restart of NS credentials are not working (I found the Problem with the Mail - IMAP Server) - after looking at the services I found sssd not running- starting the service manually fixed the Problem.

That already repeatedly happend- fortunately we´re restarting the Server quite rarely.

Could you please post the relevant part of sssd log?
Maybe we can help finding the reason.

(Mon Jul 24 23:02:34 2017) [sssd] [monitor_cleanup] (0x0010): Error removing pidfile! (2 [No such file or directory])
and
(Mon Jul 24 22:58:30 2017) [sssd[be[domain]]] [orderly_shutdown] (0x0010): SIGTERM: killing children
(Mon Jul 24 23:01:50 2017) [sssd[be[domain]]] [id_callback] (0x0010): The Monitor returned an error [org.freedesktop.DBus.Error.NoReply]
(Mon Jul 24 23:02:34 2017) [sssd[be[domain]]] [orderly_shutdown] (0x0010): SIGTERM: killing children

were domain is the DC (thye are from two different logs- one ist a sssd.log and the other the sssd_domain.log

Next Problem- related?:

Error: rpmdb open failed
Die Installation/Update ist wegen Fehlern der zwischengespeicherten Metadaten fehlgeschlagen. Bitte den Cache durch Klicken der unten stehenden Schaltfläche löschen und erneut probieren.
Any ideas!
Pushing the delete yum cache button changes nothing

Check this:

1 Like

Already found that Problem with update on two other Servers- so it´s probably a Bug with the last update.

1 Like

Thanks for the instruction- but does not solve the problem.
Further ideas and is somebody else affected?

cc/ @EddieA @GG_jr

1 Like

Any news with the problem?

Does your server still have both problems: credentials and rpmdb error?

Don’t know if it will make any difference, but you could try to increase sssd timeout:

mkdir -p /etc/e-smith/templates-custom/etc/sssd/sssd.conf/
echo 'timeout=20' > /etc/e-smith/templates-custom/etc/sssd/sssd.conf/21domain_primary
signal-event nethserver-sssd-save

The next time it fails check journalctl -u sssd.service for any error before manually starting the service.
Check also the logs for any other error.

So finaly @nikolaus.herrmann found a solution and will post that soon. Thanks a lot, for your support and inputs.

to solve the yum problen:
The best procedure to fix the RPM database is to follow the description of http://rpm.org/user_doc/db_recovery.html
I have had this phenomenon on two further systems at the same time.

It’s strange. It could be the symptom of a problem. Can you relate the problem to some activities or messages in the log files?

no, in the log files i found nothing. activity was clicking on softwarecenter