Hi,

On Wed, Feb 03, 2010 at 03:54:37AM +0200, Marian Marinov wrote:
> Hello, 
> I searched the mailing lists but didn't found answer to my problem.
> 
> I have configured a primitive with CRM but I made it a clone(the service 
> should 
> run on all nodes). So if incidentally someone start the local resource 
> instead 
> of the clone I end up with the local service and the clone have defined 
> target-
> roles and so the CRM no longer wishes to execute commands on these resources. 
> 
> Here is my configuration:
> crm(live)# configure show
> node $id="0f519a50-98e0-4bc2-8fb9-6bc28d87461c" fiona \
>         attributes standby="off"
> node $id="57db28bf-a58e-4b98-a9a9-03e65bce2431" shrek \
>         attributes standby="off"
> primitive bind lsb:named \
>         op monitor interval="5s" \
>         meta $id="bind-meta_attributes"
> primitive failover-ip ocf:heartbeat:IPaddr2 \
>         params ip="10.3.0.1" cidr_netmask="24" nic="eth0" \
>         op monitor interval="2s" \
>         meta target-role="Started"
> clone bind-clone bind \
>         meta clone-max="2" clone-node-max="1" globally-unique="false" 
> ordered="false" interleave="false" target-role="Stopped"
> location cli-standby-failover-ip failover-ip \
>         rule $id="cli-standby-rule-failover-ip" -inf: #uname eq fiona
> property $id="cib-bootstrap-options" \
>         stonith-enabled="false" \
>         last-lrm-refresh="1265161264" \
>         dc-version="1.0.6-f709c638237cdff7556cb6ab615f32826c0f8c06" \
>         cluster-infrastructure="Heartbeat"
> 
> It doesn't matter if I use OCF or LSB scripts for the test.
> 
> Here is the initial status:
> crm(live)# status
> ============
> Last updated: Wed Feb  3 03:46:30 2010
> Stack: Heartbeat
> Current DC: shrek (57db28bf-a58e-4b98-a9a9-03e65bce2431) - partition with 
> quorum
> Version: 1.0.6-f709c638237cdff7556cb6ab615f32826c0f8c06
> 2 Nodes configured, unknown expected votes
> 3 Resources configured.
> ============
> Online: [ fiona shrek ]
>  failover-ip    (ocf::heartbeat:IPaddr2):       Started shrek
> 
> Then after we start the local resource instead of the clone:
> crm(live)# resource start bind
> crm(live)# status
> ============
> Last updated: Wed Feb  3 03:46:57 2010
> Stack: Heartbeat
> Current DC: shrek (57db28bf-a58e-4b98-a9a9-03e65bce2431) - partition with 
> quorum
> Version: 1.0.6-f709c638237cdff7556cb6ab615f32826c0f8c06
> 2 Nodes configured, unknown expected votes
> 3 Resources configured.
> ============
> Online: [ fiona shrek ]
>  failover-ip    (ocf::heartbeat:IPaddr2):       Started shrek
>  Clone Set: bind-clone
>      Started: [ fiona shrek ]
> 
> The resource is working, but watch what happens when you stop the clone now:
> crm(live)# resource stop bind-clone
> Multiple attributes match name=target-role
>   Value: Stopped        (id=bind-clone-meta_attributes-target-role)
>   Value: Started        (id=bind-meta_attributes-target-role)
> Error performing operation: Required data for this CIB API call not found
> 
> Why is this automatic addition of the target-role ? What am I missing ?
> 
> My solution was:
> crm(live)# resource meta bind delete target-role
> Deleted bind option: id=bind-meta_attributes-target-role name=target-role
> crm(live)# resource stop bind-clone
> 
> But is this the right thing ?

Probably not. Though this has always been the case. crm_resource
won't touch a meta attribute if there's more than one found. You
can open an enhancement bugzilla for shell.

Thanks,

Dejan

> 
> -- Best regards, Marian Marinov



> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to