sssd services terminated

[Daniel Jung]

Hi,

Running centos 6 with sssd-1.8.0-32.el6.x86_64 and
sssd-client-1.8.0-32.el6.x86_64.

This particular problem seems to pop now and then. I looked thru the
maillist/googled and there seems to be number of posts with similar
symptoms, but didnt see exact bug number or recommended actions to resolve
this issue.
Could you advise please?  The problem resolved after restarting sssd.

sssd.log
(Tue Jul 29 16:46:06 2014) [sssd] [mt_svc_sigkill] (0x0010): [LDAP][22259]
is not responding to SIGTERM. Sending SIGKILL.
(Tue Jul 29 16:48:18 2014) [sssd] [mt_svc_sigkill] (0x0010): [nss][22262]
is not responding to SIGTERM. Sending SIGKILL.
(Tue Jul 29 16:48:18 2014) [sssd] [mt_svc_exit_handler] (0x0010): Process
[nss], definitely stopped!
(Tue Jul 29 19:01:17 2014) [sssd] [monitor_quit] (0x0010): Monitor received
Terminated: terminating children

sssd_nss.log
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_dp_init] (0x0010): Failed to
connect to monitor services.
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_process_init] (0x0010): fatal
error setting up backend connector
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_dp_init] (0x0010): Failed to
connect to monitor services.
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_process_init] (0x0010): fatal
error setting up backend connector
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_dp_init] (0x0010): Failed to
connect to monitor services.
(Tue Jul 29 16:48:18 2014) [sssd[nss]] [sss_process_init] (0x0010): fatal
error setting up backend connector

sssd_LDAP.log
(Tue Jul 29 16:48:18:940606 2014) [sssd[be[LDAP]]]
[confdb_get_domain_internal] (0x0020): No enumeration for [LDAP]!
(Tue Jul 29 16:48:19:181330 2014) [sssd[be[LDAP]]] [be_process_init]
(0x0020): No Session module provided for [LDAP] !!
(Tue Jul 29 16:48:19:181468 2014) [sssd[be[LDAP]]] [be_process_init]
(0x0020): No host info module provided for [LDAP] !!
(Tue Jul 29 16:48:19:181550 2014) [sssd[be[LDAP]]] [main] (0x0020): Backend
provider (LDAP) started!
(Tue Jul 29 19:01:17:891413 2014) [sssd[be[LDAP]]]
[confdb_get_domain_internal] (0x0020): No enumeration for [LDAP]!
(Tue Jul 29 19:01:17:934694 2014) [sssd[be[LDAP]]] [be_process_init]
(0x0020): No Session module provided for [LDAP] !!
(Tue Jul 29 19:01:17:934781 2014) [sssd[be[LDAP]]] [be_process_init]
(0x0020): No host info module provided for [LDAP] !!
(Tue Jul 29 19:01:17:934831 2014) [sssd[be[LDAP]]] [main] (0x0020): Backend
provider (LDAP) started!

[Jakub Hrozek]

On Tue, Jul 29, 2014 at 11:07:37AM -0700, Daniel Jung wrote:
> Hi,
>
> Running centos 6 with sssd-1.8.0-32.el6.x86_64 and
> sssd-client-1.8.0-32.el6.x86_64.

These are very old versions, first recommendation is to upgrade. In 6.5,
we have a heavily patched 1.9 branch (pretty much equivalent to 1.9.6
upstream).
> This particular problem seems to pop now and then. I looked thru the
> maillist/googled and there seems to be number of posts with similar
> symptoms, but didnt see exact bug number or recommended actions to resolve
> this issue.
> Could you advise please?  The problem resolved after restarting sssd.
>
> sssd.log
> (Tue Jul 29 16:46:06 2014) [sssd] [mt_svc_sigkill] (0x0010): [LDAP][22259]
> is not responding to SIGTERM. Sending SIGKILL.

I suspect you're running with enumerate=true ? If so, then I suspect
your backend took too long to save all entries and was killed by the
monitor. Either disable enumeration (preferred) or increase the
'timeout' parameter a bit.

[Daniel Jung]

Hi Jakub,

Upgrading to the version avail in 6.5 is  certainly an option. Is similar
package avail for centos5? i see that1.5.1 LTM is what's being avail.
We have enumeration disabled.

On Wed, Jul 30, 2014 at 1:48 AM, Jakub Hrozek  wrote:
> On Tue, Jul 29, 2014 at 11:07:37AM -0700, Daniel Jung wrote:
> > Hi,
> >
> > Running centos 6 with sssd-1.8.0-32.el6.x86_64 and
> > sssd-client-1.8.0-32.el6.x86_64.
>
> These are very old versions, first recommendation is to upgrade. In 6.5,
> we have a heavily patched 1.9 branch (pretty much equivalent to 1.9.6
> upstream).
> > This particular problem seems to pop now and then. I looked thru the
> > maillist/googled and there seems to be number of posts with similar
> > symptoms, but didnt see exact bug number or recommended actions to
>
> resolve
> > this issue.
> > Could you advise please?  The problem resolved after restarting sssd.
> >
> > sssd.log
> > (Tue Jul 29 16:46:06 2014) [sssd] [mt_svc_sigkill] (0x0010):
>
> [LDAP][22259]
> > is not responding to SIGTERM. Sending SIGKILL.
>
> I suspect you're running with enumerate=true ? If so, then I suspect
> your backend took too long to save all entries and was killed by the
> monitor. Either disable enumeration (preferred) or increase the
> 'timeout' parameter a bit.
> _______________________________________________
> sssd-users mailing list
> [e-mail]
> https://lists.fedorahosted.org/mailman/listinfo/sssd-users

[Jakub Hrozek]

On Wed, Jul 30, 2014 at 10:01:57AM -0700, Daniel Jung wrote:
> Hi Jakub,
>
> Upgrading to the version avail in 6.5 is  certainly an option. Is similar
> package avail for centos5? i see that1.5.1 LTM is what's being avail.
> We have enumeration disabled.

No, el5 is stuck on 1.5.x. You can get (unsupported, only upstream
tested) 1.9 packages in Stephen's repo:
    http://copr-fe.cloud.fedoraproject.org/coprs/sgallagh/sssd-1.9-rhel5/