[Linux-HA] Problem with multiple drbd ressources in master_slave section

[Andrew Beekhof]

On Thu, Apr 17, 2008 at 10:46 AM, Nicolas Castel
<castelnicolas at gmail.com> wrote:
> Hi all,
>
>  I use heartbeat-2.1.3-3.el5.centos on RHEL5 and drbd-8.0.11 built from
>  sources.
>
>  I try to deploy 3 drbd ressources inside a group in master_slave section.

I'd not be doing that with 2.1.3... you'll want the upcoming 0.6.3
version of pacemaker (early next week) as there have been some bugs
fixed in this area.

>  With only one drbd ressource in master_slave section, the configuration
>  works fine (without group in master_slave section).
>
>  With 3 drbd ressources, the first one starts correctly, but the second fails
>  and so heartbeat does not try to start the third.

Looks like an RA issue... the myserver01-drbd-log is returning stopped
in its monitor action.
Are you sure it's able to function as a master/slave resource?

>
>  Here is the output of crm_mon command.
>
>  ============
>  Last updated: Thu Apr 17 09:09:20 2008
>  Current DC: sharav (cdb1a5d7-31c8-477d-9319-8205d268eb1c)
>  2 Nodes configured.
>  2 Resources configured.
>  ============
>
>  Node: sharav (cdb1a5d7-31c8-477d-9319-8205d268eb1c): online
>  Node: shamal (4ed1d0b3-840f-4ca8-8f18-9515155ab0b5): online
>
>  Master/Slave Set: ms-myserver01-drbd
>     Resource Group: rg-drbd:0
>         myserver01-drbd-data:0  (heartbeat::ocf:drbd):  Started shamal
>         myserver01-drbd-log:0   (heartbeat::ocf:drbd):  Started shamal
>  FAILED
>         myserver01-drbd-tmp:0   (heartbeat::ocf:drbd):  Stopped
>     Resource Group: rg-drbd:1
>         myserver01-drbd-data:1  (heartbeat::ocf:drbd):  Started sharav
>         myserver01-drbd-log:1   (heartbeat::ocf:drbd):  Started sharav
>  FAILED
>         myserver01-drbd-tmp:1   (heartbeat::ocf:drbd):  Stopped
>
>  Failed actions:
>     myserver01-drbd-log:1_monitor_30000 (node=sharav, call=41, rc=7):
>  complete
>     myserver01-drbd-log:0_monitor_30000 (node=shamal, call=12, rc=7):
>  complete
>
>  Moreover, in this case, heartbeat service can't be stopped.
>
>  Here is my master_slave configuration section :
>
>        <master_slave id="ms-myserver01-drbd">
>          <meta_attributes id="ma-ms-myserver01-drbd">
>            <attributes>
>              <nvpair id="ma-ms-myserver01-drbd-1" name="clone_max"
>  value="2"/>
>              <nvpair id="ma-ms-myserver01-drbd-2" name="clone_node_max"
>  value="1"/>
>              <nvpair id="ma-ms-myserver01-drbd-3" name="master_max"
>  value="1"/>
>              <nvpair id="ma-ms-myserver01-drbd-4" name="master_node_max"
>  value="1"/>
>              <nvpair id="ma-ms-myserver01-drbd-5" name="notify"
>  value="yes"/>
>              <nvpair id="ma-ms-myserver01-drbd-6" name="globally_unique"
>  value="false"/>
>              <nvpair id="ma-ms-myserver01-drbd-7" name="target_role"
>  value="#default"/>
>            </attributes>
>          </meta_attributes>
>          <group id="rg-drbd">
>            <primitive id="myserver01-drbd-data" class="ocf"
>  provider="heartbeat" type="drbd">
>              <instance_attributes id="ia-myserver01-drbd-data">
>                <attributes>
>                  <nvpair id="ia-myserver01-drbd-data-1" name="drbd_resource"
>  value="myserver01-data"/>
>                </attributes>
>              </instance_attributes>
>              <operations>
>                <op id="op-myserver01-drbd-data-1" name="monitor"
>  interval="29s" timeout="15s" role="Master"/>
>                <op id="op-myserver01-drbd-data-2" name="monitor"
>  interval="30s" timeout="15s" role="Slave"/>
>              </operations>
>            </primitive>
>            <primitive id="myserver01-drbd-log" class="ocf"
>  provider="heartbeat" type="drbd">
>              <instance_attributes id="ia-myserver01-drbd-log">
>                <attributes>
>                  <nvpair id="ia-myserver01-drbd-log-1" name="drbd_resource"
>  value="myserver01-log"/>
>                </attributes>
>              </instance_attributes>
>              <operations>
>                <op id="op-myserver01-drbd-log-1" name="monitor"
>  interval="29s" timeout="15s" role="Master"/>
>                <op id="op-myserver01-drbd-log-2" name="monitor"
>  interval="30s" timeout="15s" role="Slave"/>
>              </operations>
>            </primitive>
>            <primitive id="myserver01-drbd-tmp" class="ocf"
>  provider="heartbeat" type="drbd">
>              <instance_attributes id="ia-myserver01-drbd-tmp">
>                <attributes>
>                  <nvpair id="ia-myserver01-drbd-tmp-1" name="drbd_resource"
>  value="myserver01-tmp"/>
>                </attributes>
>              </instance_attributes>
>              <operations>
>                <op id="op-myserver01-drbd-tmp-1" name="monitor"
>  interval="29s" timeout="15s" role="Master"/>
>                <op id="op-myserver01-drbd-tmp-2" name="monitor"
>  interval="30s" timeout="15s" role="Slave"/>
>              </operations>
>            </primitive>
>          </group>
>        </master_slave>
>
>  What wrong ?
>
>  Thanks in advance.
>  _______________________________________________
>  Linux-HA mailing list
>  Linux-HA at lists.linux-ha.org
>  http://lists.linux-ha.org/mailman/listinfo/linux-ha
>  See also: http://linux-ha.org/ReportingProblems
>