Hi all,

I'm new to the Heartbeat. I'm running 2.0.8. My cluster configuration is as 
follows:
- 2 nodes.
- I've configured a OCF DRBD master/slave resource.
- OCF Filesystem resource which starts on the master node.
- A specific service runs on the node where the filesystem resource is mounted. 
This service runs with /etc/init.d script therefore its class is LSB.

What I want to configure, and I don't know how is when the service fails to 
promote the other(second) node DRBD resource to master, mount the filesystem 
and run the resource on the second node. Very common situation - if service 
fails on the first node to failover it to the second. This is very easy to 
configure with <op id="1" name="monitor" interval="10s" timeout="4s"/>. The 
heartbeat monitors to service. If it fails, heartbeat reallocate the resource 
to the other node.

However, the tricky part here is the fact that the service needs the DRBD 
resource promoted and filesystem mounted on the second node. How to move all 
resources on the other node if my service fails?

I can define filesystem and the service in a groups - this will move both them 
on the second node. But I can't put DRBD resource in the group because it is 
Master/Slave type.

So my question is how to promote the DBRD resource on the other node when my 
service fails? What constraints should I set? Browsing the web I wasn't able to 
find anything on this matter. 

How to define an expression which says: "If my service has failed decrease the 
score for role=master for this node."? or increase to score for slave...

How to trigger an action if certain service has failed?

Any help would be appreciated a lot.

Thanks in advance.
Atanas Dyulgerov

-----------------------------------------------------------------
Оцеляваш ли? Сподели свое видео или снимки! http://robinson.btv.bg/isurvive
_______________________________________________
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