[Linux-cluster] altname broken?

[Frederik Wagner]

Hi Patrick,

On 06/13/2007 11:09 AM, Patrick Caulfield wrote:
> Frederik Wagner wrote:
>>
>> Are there any known problems regading this option?
>>
> 
> There's nothing changed in that area that I can recall. Can you track the
> startup with "cman_tool join -d" and also CMAN messages that appear in syslog on
> the nodes?
> 

I reintroduced the altname in my cluster.conf and get the following problem.
Starting the service "cman" on the three cluster nodes makes then hang
while trying to start the fencing:

# service cman start
Starting cluster:
   Loading modules... done
   Mounting configfs... done
   Starting ccsd... done
   Starting cman... done
   Starting daemons... done
   Starting fencing... failed

                                                           [FAILED]

after some minutes.
While stoping the script, and running clustat shows, that the node only
sees itself as Online, but not the others

# clustat
Member Status: Inquorate

  Member Name                        ID   Status
  ------ ----                        ---- ------
  auriga                                1 Offline
  pod                                   2 Offline
  sulaco                                3 Online, Local

also ccs_tool doesn't  see the other nodes (e.g. ccs_tool update... does
not work with "Unable to connect to the CCS daemon: Connection refused").

The output of the cman_tool join -d (after starting ccsd by hand and
loading the modules, i suppose) is attached as output.txt. I left it a
bit running while calling clustat etc...

The relevant lines of my cluster.conf are the following:

<?xml version="1.0"?>
<cluster config_version="30" name="storage_cluster">
 <fence_daemon post_fail_delay="0" post_join_delay="3"/>
 <clusternodes>
  <clusternode name="auriga" nodeid="1" votes="1">
   <altname name="auriga-hb"/>
   <fence>
    <method name="ipmi">
     <device lanplus="1" name="auriga_IPMI"/>
    </method>
    <method name="fabric">
     <device name="fcswitch1" port="5"/>
     <device name="fcswitch2" port="5"/>
    </method>
   </fence>
  </clusternode>
  <clusternode name="pod" nodeid="2" votes="1">
   <altname name="pod-hb"/>
   <fence>
    <method name="ipmi">
     <device lanplus="1" name="pod_IPMI"/>
    </method>
    <method name="fabric">
     <device name="fcswitch1" port="6"/>
     <device name="fcswitch2" port="6"/>
    </method>
   </fence>
  </clusternode>
  <clusternode name="sulaco" nodeid="3" votes="1">
   <altname name="sulaco-hb"/>
   <fence>
    <method name="ipmi">
     <device lanplus="1" name="sulaco_IPMI"/>
    </method>
    <method name="fabric">
     <device name="fcswitch1" port="4"/>
     <device name="fcswitch2" port="4"/>
    </method>
   </fence>
  </clusternode>
 </clusternodes>
 <cman/>
 <fencedevices>
  ...
 </fencedevices>
 <rm>
  ...
 </rm>
</cluster>

Thanks for any help! And let me know if you need further output...

Frederik
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: output.txt
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20070613/a71e9cd5/attachment.txt>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: OpenPGP digital signature
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20070613/a71e9cd5/attachment.sig>