Created attachment 863615 [details]
sssd.log

During update, sssd.rpm tries to restart all sssd sevices and sockets, even if
they cannot be restarted. 

This leads to a lot of error messages and sometimes a non functional sssd
service. See the sssd.log in the attachment. According to the admin, it happens
on a regular base that they need to restart the daemon.

Please restart only this services in the correct order, which lead to a
functional system afterwards! Don't restart *.socket services, it's really not
necessary. And don't try to restart services which cannot be restarted.
If rpmlint complains, ignore that or add an exception.
Prior 1 has to be that the services continues to work after an update and the
admin does not need to fix it afterwards.
If it is not possible to restart sssd in a secure, safe way, better don't do so
before you lockout the admin.

In this case sssd even seg.faulted in libsss_utils.so, no idea why only in this
place and not always. Maybe other, depending libraries got only update
afterwards and not before? If this can happen, restarting a service in a
pre/post install script is even more dangerous.