Alessandro,
this second log starts to show something.
So first one:
Feb 15 21:27:43 corosync [TOTEM ] Process pause detected for 17920 ms,
flushing membership messages.
Feb 15 21:27:43 corosync [MAIN ] Corosync main process was not
scheduled for 18256.9629 ms (threshold is 2400.0000 ms). Consid
er token timeout increase.
VM not scheduled for 18 sec? That's A LOT. Please try to first solve
this issue.
Also because you are running two node, it may make sense to have <cman
two_node="1" expected_votes="1"/> snip in configuration file (see
https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Cluster_Administration/s1-creating-cluster-cli-CA.html).
I will try to reproduce issue you are hitting, because error 2
shouldn't happen.
Regards,
Honza
Alessandro Bono napsal(a):
Il 17/02/14 10:29, Jan Friesse ha scritto:
Alessandro,
can you please send me also log from second node?
sure
I'll send as private email, log file compressed is 420kb
But honestly, log seems pretty weird. I mean, there were token lost,
single node formation, join of second node an d then this library
error. Similar problem was in rhel 6.4 package but I believed it is
properly fixed.
Regards,
Honza
Alessandro Bono napsal(a):
Hi Honza
attached new log with corosync "Library error"
this time with this configuration
<cluster config_version="8" name="ga-ext_cluster">
<totem token="3000" consensus="5000" />
<logging>
<logging_daemon name="corosync" debug="on"/>
</logging>
<clusternodes>
<clusternode name="ga1-ext" nodeid="1">
<fence>
<method name="pcmk-redirect">
<device name="pcmk" port="ga1-ext"/>
</method>
</fence>
</clusternode>
<clusternode name="ga2-ext" nodeid="2">
<fence>
<method name="pcmk-redirect">
<device name="pcmk" port="ga2-ext"/>
</method>
</fence>
</clusternode>
</clusternodes>
<fencedevices>
<fencedevice agent="fence_pcmk" name="pcmk"/>
</fencedevices>
</cluster>
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
