[Pacemaker] How to set up logging for CTSlab

[Andrew Beekhof]

On Mon, Jan 17, 2011 at 3:26 PM, Simon Jansen
<simon.jansen1 at googlemail.com> wrote:
>> On Mon, Jan 10, 2011 at 5:40 PM, Dejan Muhamedagic <deja... at fastmail.fm>
>> wrote:
>> > Hi,
>> >
>> > On Mon, Jan 10, 2011 at 10:46:58AM +0100, Simon Jansen wrote:
>> >> Hi,
>> >>
>> >> I would like to test my cluster with the cluster test suite. I followed
>> >> the
>> >> instructions of the README file to set up the logging with syslog-ng.
>> >> Further I changed the syslog_facility in the corosync.conf file to
>> >> local7.
>> >> The cluster nodes are sending their logs to the exerciser and the
>> >> exerciser
>> >> writes the log entries into the file /var/log/ha-log-local7. The
>> >> entries
>> >> have the form
>> >> Jan 10 10:27:11 192.168.200.10 corosync[903]:   [SERV  ] Unloading all
>> >> Corosync service engines.
>> >> Jan 10 10:27:11 192.168.200.10 corosync[903]:   [pcmk  ] notice:
>> >> pcmk_shutdown: Shuting down Pacemaker
>> >> Jan 10 10:27:11 192.168.200.10 corosync[903]:   [pcmk  ] notice:
>> >> pcmk_shutdown: mgmtd confirmed stopped
>> >> Jan 10 10:27:11 192.168.200.10 corosync[903]:   [pcmk  ] notice:
>> >> stop_child:
>> >> Sent -15 to crmd: [1023]
>> >> So I would say that the remote logging is working.
>> >> But when I start the CTS by executing the following command on the
>> >> exerciser
>> >>
>> >>   python CTSlab.py --nodes "192.168.200.10 192.168.200.20" --at-boot 1
>> >> --stack corosync --stonith no --logfile /var/log/ha-log-local7 1
>> >> the LogAudit fails.
>> >>
>> >> The output of CTS is
>> >> Jan 10 10:04:03 Random seed is: 1294650243
>> >> Jan 10 10:04:03 >>>>>>>>>>>>>>>> BEGINNING 1 TESTS
>> >> Jan 10 10:04:03 Stack:            corosync (flatiron)
>> >> Jan 10 10:04:03 Schema:           pacemaker-1.0
>> >> Jan 10 10:04:03 Scenario:         Random Test Execution
>> >> Jan 10 10:04:03 Random Seed:      1294650243
>> >> Jan 10 10:04:03 System log files: /var/log/ha-log-local7
>> >> Jan 10 10:04:03 Cluster nodes:
>> >> Jan 10 10:04:03 * 192.168.200.10
>> >> Jan 10 10:04:03 * 192.168.200.20
>> >> Jan 10 10:06:17 Testing for syslog logs
>> >> Jan 10 10:06:17 Testing for remote logs
>> >> Jan 10 10:07:26 Restarting logging on: ['192.168.200.10',
>> >> '192.168.200.20']
>> >> Jan 10 10:09:01 Restarting logging on: ['192.168.200.10',
>> >> '192.168.200.20']
>> >> Jan 10 10:11:37 Restarting logging on: ['192.168.200.10',
>> >> '192.168.200.20']
>> >> Jan 10 10:15:02 Restarting logging on: ['192.168.200.10',
>> >> '192.168.200.20']
>> >> Jan 10 10:19:02 ERROR: Cluster logging unrecoverable.
>> >> Jan 10 10:19:02 Audit LogAudit FAILED.
>> >
>> > What happened to the node names? I mean, your nodes are being
>> > named after their IP addresses. That's unusual.
>> >
>
> The exerciser machine is placed in another network than the public one and I
> didn't defiend the hosts in the /etc/hosts file. The IP addresses are in the
> log summary on the exercise, too. So I thought the IP would be the best
> choice.
>
>>
>> > Otherwise, it should work. BTW, what do you see when you run
>> > "logger -p local7.info hi" on one of the nodes? Does that end in
>> > the syslog on your central host?
>
>  Yes, it does end in the syslog of the exerciser.
> Today I tested the following command:
> python CTSlab.py --nodes "192.168.200.10 192.168.200.20" --logfile
> /var/log/ha-log-local7 --syslog-facility local7 --no-unsafe-tests --at-boot
> 1
> --stack corosync --stonith no 1
>
>>
>> thats not a requirement anymore
>
> If it is not a requirement anymore, how do I have to call the CTSlab script
> and what parameters do I need?

Its auto-detected.
The only requirement is that the values for --nodes must match the
result of uname -n for the cluster nodes.

So perhaps adding them to /etc/hosts is the simplest path forward.

>
> Thanks for your help.
> --
>
>
> Best regards,
>
> Simon Jansen
>
>
> ---------------------------
> Simon Jansen
> 64291 Darmstadt
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs:
> http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
>
>