Are sssd_dp errors worth reporting?

[Jeff Schroeder]

I've seen an occasional and (as of yet) impossible to reproduce dp
segfault with sssd 0.6-1. Would it be worth trying to track down? It
looked like there were patches merged to remove the dp process
entirely. When sssd_dp segfaults, the watchdog doesn't restart it and
bad things happen. That leads to nasty messages in syslog like:

Oct  9 18:20:01.122 nameobscured crond[6931]: pam_sss(crond:account):
Request to sssd failed.

Restarting sssd fixes the problem but this is a bit disconcerting. Is
master in shape to snapshot and build rpms from or should I wait for
the 0.7 release without sssd_dp? This is the second time we've seen
this since 0.6 was released but it is still a legit problem.

-- 
Jeff Schroeder

Don't drink and derive, alcohol and analysis don't mix.
http://www.digitalprognosis.com

[Stephen Gallagher]

Jeff, if you're seeing sssd_dp segfaults, can you get us a core dump? We're
planning to remove the sssd_dp in the 0.7.0 release in a few weeks, but
we're also going to be doing a 0.6.1 maintenance release (that we planned to
release on Monday)
If you can get us a core dump (or at least a backtrace), I'll see what I can
do about fixing it for the 0.6.1 release.

On Fri, Oct 9, 2009 at 6:50 PM, Jeff Schroeder <jeffschroed(a)gmail.com&gt;wrote:

>  I've seen an occasional and (as of yet) impossible to reproduce
> dp
>  segfault with sssd 0.6-1. Would it be worth trying to track down? It
>  looked like there were patches merged to remove the dp process
>  entirely. When sssd_dp segfaults, the watchdog doesn't restart it and
>  bad things happen. That leads to nasty messages in syslog like:
>
>  Oct  9 18:20:01.122 nameobscured crond[6931]: pam_sss(crond:account):
>  Request to sssd failed.
>
>  Restarting sssd fixes the problem but this is a bit disconcerting. Is
>  master in shape to snapshot and build rpms from or should I wait for
>  the 0.7 release without sssd_dp? This is the second time we've seen
>  this since 0.6 was released but it is still a legit problem.
>
>  --
>  Jeff Schroeder
>
>  Don't drink and derive, alcohol and analysis don't mix.
>  http://www.digitalprognosis.com
>  _______________________________________________
>  sssd-devel mailing list
>  sssd-devel(a)lists.fedorahosted.org
>  https://fedorahosted.org/mailman/listinfo/sssd-devel
>