[Freeipa-users] sssd stops after nss crashes

[Lukas Slebodnik]

On (12/09/16 11:09), Lachlan Musicman wrote:
>We saw another sssd crash on the weekend (well, Friday night).
>
>Centos 7, sssd 1.14.0 from COPR
>
Please upgrade to 1.14.1 from copr.

>Everything has worked fine for over a month until Friday.
>
>According to the log sssd_nss on the host in question:
>
> - at about 16:18, watchdog_handler killed a process for a timer overflow.
> - there is some flopping about as nss/sssd tries to reconnect
> - at 16:19:12 we see this:
>
>(Fri Sep  9 16:19:12 2016) [sssd[nss]] [sbus_dispatch] (0x0400): SBUS is
>reconnecting. Deferring.
>
> - Which continues until 16:20:56
>
>(Fri Sep  9 16:20:56 2016) [sssd[nss]] [sbus_dispatch] (0x0400): SBUS is
>reconnecting. Deferring.
>
>Note that there are 9,573,091 lines of this, at about 80,000 msgs per
>second.
>
> - nss seems to stumble back to life at this point (there are no logs on
>the freeipa server unfortunately)
>
> - at every 15 min interval we see this (I think this might be zabbix
>polling sssd):
>
>(Fri Sep  9 18:30:01 2016) [sssd[nss]] [get_client_cred] (0x0020):
>SELINUX_getpeercon failed [-1][Unknown error -1].
What is a state of SELinux on your machine?
Please share output of "sestatus"

LS