[Pacemaker] CRM and openAIS
Andrew Beekhof
beekhof at gmail.com
Fri Jan 11 06:33:26 UTC 2008
On Jan 10, 2008, at 10:28 PM, Serge Dubrouski wrote:
> Ok.
>
> 1. I've got Fedora Core 6 packages from you site and installed them.
> 2. While installing pacemaker-openais complained on conflict over
> /usr/lib/heartbeat/haresources2cib.py file that was installed with
> heartbeat-common and again included into pacemaker-openais
ah, ok. i'll fix that
>
> 3. openais and pacemaker-openasi rpms produced following non critical
> error message during installation:
>
> /var/tmp/rpm-tmp.NNNNNN: line 1: fg: no job control
Hmmm... i'll investigate
>
>
> 4. For some yet unknown reason I wasn't able to make openais to log
> into /var/log/openais.log. My logging section of config file looks
> like this:
>
> logging {
> fileline: off
> to_stderr: yes
> to_file: yes
> to_syslog: yes
> syslog_facility: daemon
> logfile: /var/log/openais.log
> debug: on
> timestamp: on
> logger_subsys {
> subsys: CRM
> debug: off
> tags: enter|leave|trace1|trace2|trace3|trace4|trace6
> fileline: off
> }
> }
i think (you might need to check with the openais guys) that only one
type of logging is performed and syslog_facility takes precedent over
logfile
>
>
> 5. openais started all right but CRM failed to start:
>
> Jan 10 16:13:01 fc-node2 openais[16305]: [SERV ] Service initialized
> 'openais configuration service'
> Jan 10 16:13:01 fc-node2 openais[16305]: [SERV ] Service initialized
> 'openais cluster closed process group service v1.01'
> Jan 10 16:13:01 fc-node2 openais[16305]: [SERV ] Service failed to
> load 'lha_crm'.
> Jan 10 16:13:01 fc-node2 openais[16305]: [SYNC ] Not using a virtual
> synchrony filter.
> Jan 10 16:13:01 fc-node2 openais[16305]: [TOTEM] Creating commit
> token because I
>
> I wasn't able to find any additional information.
can you do an ls of /usr/lib/lcrso for me pls?
looks like the plugin cant be found or wont load
> 6. Looks like openias itself is healthy. It reported both nodes
> joining the cluster:
>
> Jan 10 16:13:31 fc-node2 openais[16305]: [CLM ] CLM CONFIGURATION
> CHANGE
> Jan 10 16:13:31 fc-node2 openais[16305]: [CLM ] New Configuration:
> Jan 10 16:13:31 fc-node2 openais[16305]: [CLM ] r(0)
> ip(192.168.1.111)
> Jan 10 16:13:31 fc-node2 openais[16305]: [CLM ] r(0)
> ip(192.168.1.112)
>
> 7. None of the installed packages included ocf RAs. How is it supposed
> to run resources?
they're in heartbeat-resources
though i'm not 100% sure breaking them out into their own package is a
good idea. thoughts?
>
>
> Thanks.
>
> On Jan 10, 2008 10:52 AM, Andrew Beekhof <beekhof at gmail.com> wrote:
>>
>>
>> On Jan 10, 2008, at 5:43 PM, Serge Dubrouski wrote:
>>
>>> Hello -
>>>
>>> So I've beem looking for some documentation how to marry CRM and
>>> openAIS and so far found none. I suppose that openAIS should support
>>> some kind of crm {} section in openais.conf file like it does for
>>> amf,
>>> is it true? Are there any hints on how to make openAIS to start CRM?
>>
>> i'll have to write a howto on this when the official packages get
>> released next week, but for now:
>>
>> 1) you need to get the openais^, pacemaker-ais and heartbeat-common
>> packages from http://download.opensuse.org/repositories/home:/
>> beekhof/
>> 2) set up a regular openais configuration. eg.
>> totem {
>> version: 2
>> secauth: off
>> threads: 0
>> interface {
>> ringnumber: 0
>> bindnetaddr: 192.168.9.51
>> mcastaddr: 226.94.1.1
>> mcastport: 5405
>> }
>> }
>>
>> logging {
>> fileline: off
>> to_syslog: yes
>> syslog_facility: daemon
>> to_stderr: yes
>> to_file: yes
>> logfile: /var/log/openais.log
>> debug: on
>> timestamp: on
>> logger {
>> ident: CRM
>> debug: on
>> tags: enter|leave|trace1|trace2|trace3|trace4|trace6
>> fileline: off
>> }
>> }
>>
>> amf {
>> mode: disabled
>> }
>>
>> 3) add the following to /etc/ais/openais.conf
>> aisexec {
>> user: root
>> group: root
>> }
>>
>> 4) run: /etc/init.d/ais start
>>
>> 5) repeat on the rest of your nodes.
>>
>> 6) thats it. the same cib configuration can be used. um, except for
>> the nodes section... probably best to wipe that (since the UUIDs wont
>> be the same)
>>
>> 7) send feedback :-)
>>
>>
>> ^ only the packages at that location have the two extra patches (one
>> from red hat, one from me) required for running pacemaker
>>
>>>
>>>
>>> Thanks.
>>>
>>> --
>>> Serge Dubrouski.
>>>
>>> _______________________________________________
>>> Pacemaker mailing list
>>> Pacemaker at clusterlabs.org
>>> http://list.clusterlabs.org/mailman/listinfo/pacemaker
>>
>>
>
>
>
> --
> Serge Dubrouski.
More information about the Pacemaker
mailing list