[Pacemaker] Master/Slave drbd setup not working as expected

Jake Smith jsmith at argotec.com
Mon Sep 10 13:19:05 EDT 2012


----- Original Message -----
> From: "Luca Meron" <ckslxpcjus at hotmail.it>
> To: "The Pacemaker cluster resource manager" <pacemaker at oss.clusterlabs.org>
> Sent: Monday, September 10, 2012 12:39:30 PM
> Subject: [Pacemaker] Master/Slave drbd setup not working as expected
> 
> Hi.
> I've a simple config with Corosync 1.4.2 on Ubuntu 12.04 which is
> causing me some troubles.
> I'm actually handling two drbd master/slave resources. The problem is
> that one of the two is not activated on the slave, causing it to
> remain out of sync!
> The setup is very simple (follows), the Config master loads
> correctly, while the Firewall one doesn't! Becomes primary on
> master, but on slave remains stopped.
> What am I doing wrong?
> 
> thanks!
> 
> 
> # crm_mon -1
> ============
> Last updated: Mon Sep 10 18:39:03 2012
> Last change: Mon Sep 10 18:36:39 2012 via crm_attribute on fucking
> Stack: openais
> Current DC: deposto - partition with quorum
> Version: 1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c
> 2 Nodes configured, 2 expected votes
> 6 Resources configured.
> ============
> 
> Online: [ deposto fucking ]
> 
>  Master/Slave Set: Config [ConfigDrbd]
>      Masters: [ deposto ]
>      Slaves: [ fucking ]
>  Master/Slave Set: Firewall [FirewallDrbd]
>      Masters: [ deposto ]
>      Stopped: [ FirewallDrbd:1 ]
>  ConfigDir      (ocf::heartbeat:Filesystem):    Started deposto
>  FirewallVM     (ocf::heartbeat:VirtualDomain): Started deposto
> 
> 
> config:
> 
> node deposto
> node fucking
> primitive ConfigDir ocf:heartbeat:Filesystem \
>         params device="/dev/drbd/by-res/config" directory="/config"
>         fstype="ext4"
> primitive ConfigDrbd ocf:linbit:drbd \
>         params drbd_resource="config" \
>         op monitor interval="90s"
> primitive FirewallDrbd ocf:linbit:drbd \
>         params drbd_resource="firewall" \
>         op monitor interval="90s"
> primitive FirewallVM ocf:heartbeat:VirtualDomain \
>         params config="/config/fw.kvm" \
>         op monitor interval="60s"
> ms Config ConfigDrbd \
>         meta master-max="1" master-node-max="1" clone-max="2"
>         clone-node-max="1" notify="true" globally-unique="false"
> ms Firewall FirewallDrbd \
>         meta master-max="1" master-node-max="1" clone-max="2"
>         clone-node-max="1" notify="true" globally-unique="false"
> colocation ConfigDirWithWork inf: ConfigDir Config:Master

This one could be the problem:
> colocation FirewallDiskWithWork inf: Firewall Config:Master

I believe that is only allowing Firewall to run where Config:Master runs.  Without specifying the role in the collocation the Firewall:slave cannot run on a different node.  I assume you are trying to have the DRBD masters on the same node?  That should be something like:

colocation FirewallDiskWithWork inf: Firewall:Master Config:Master

If that doesn't fix it then:

Were both DRBD resources working as expected before you introduced pacemaker?
Output of $cat /proc/drbd on both nodes?
Logs when the drbd resources are started from both nodes?  They should hold the key to the problem.

HTH

Jake

> colocation FirewallVMWithFirewallDisk inf: FirewallVM Firewall:Master
> order ConfigBeforeConfigDir inf: Config:promote ConfigDir:start
> order FirewallDiskBeforeFirewallVM inf: Firewall:promote
> FirewallVM:start
> property $id="cib-bootstrap-options" \
>         dc-version="1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c" \
>         cluster-infrastructure="openais" \
>         expected-quorum-votes="2" \
>         no-quorum-policy="ignore" \
>         stonith-enabled="false" \
>         default-resource-stickiness="1000" \
>         last-lrm-refresh="1347293917"
> 
> 
> --
> 
> 
> _______________________________________________
> 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://bugs.clusterlabs.org
> 
> 




More information about the Pacemaker mailing list