florian at hastexo
Feb 29, 2012, 8:43 AM
Post #7 of 10
On Wed, Feb 29, 2012 at 5:14 PM, Marcus Bointon
<marcus [at] synchromedia> wrote:
> On 29 Feb 2012, at 16:33, Florian Haas wrote:
>> No, there's an easier way to fix that problem. :)
>> You said this was a vanilla config that needn't be preserved, right?
>> Shut down Corosync on both nodes. Kill the contents of
>> /var/lib/heartbeat/crm/. Then bring everything back up.
> That worked fine on www5 but not www4, which didn't recreate the cib files. This time though it did not log any errors, all looks reasonable, but crm status is still failing to connect, there's still no cib process, and now www5 can't seem to see it either. I tried copying over the cib files from www5 (which seemed to be an empty xml config) but it didn't help: cib still isn't running.
> Also now www5 no longer finds itself - crm status reports 0 nodes.
My hunch is that you never properly shut down corosync on that one.
Did you check your ps output so see if it was really down? Corosync
1.2.x had some nasty shutdown issues when running with Pacemaker.
Let us know if that helped. Thanks!
Need help with High Availability?
Linux-HA mailing list
Linux-HA [at] lists
See also: http://linux-ha.org/ReportingProblems