[ClusterLabs] Q: (SLES11 SP4) Truncated lines in /var/log/pacemaker.log?

Ken Gaillot kgaillot at redhat.com
Thu Nov 3 15:22:44 CET 2016


On 11/03/2016 04:31 AM, Ulrich Windl wrote:
> Hi!
> 
> Looking at /var/log/pacemaker.log I got the impression that lines are truncated at 512 bytes. Is that true, and if so: Why? Pacemaker creates lines longer than that limit.
> For example a "cib:     info: cib_perform_op:  ++ /cib/status/node_state" usually exceeds that limit.

Not all messages are truncated. One of our XML loggers truncates long
XML lines to avoid flooding the logs. In those logs, the XML is just a
bit of extra debugging detail to get an idea of where it is, so nothing
is really lost.

> Following some changes I had made, I also found out that when changing the monitoring interval for a resource pacemaker tries to look up a device with that name for stonith. Why that?
> stonith-ng:     info: stonith_device_remove:   Device 'prm_t11_ascs_ers' not found (2 active devices)

Yes, that annoys me every time I see it. :) It's at info level to
indicate it's not a problem of any kind. I forget the exact details, but
stonithd checks CIB updates for any resource to see whether they affect
any registered fence devices.



More information about the Users mailing list