Fail over in Linux cluster - Active/Passive

[khan_dba]

Dear Experts,

I am an Oracle DBA and we got a Linux cluster setup in Active/Passive mode. I have defined a resource script which is to start the Database and listener service during failover. The script is getting run every minute and trying to start the database again and again even though the database has been started already. Is this a general behavior of the Linux cluster? What do I need to perform so that my script is invoked only once? Any assistance will be highly appreciated.

Thanks in Advance!
Khan.

[khan_dba]

Dear Experts,

I am an Oracle DBA and we got a Linux cluster setup in Active/Passive mode. I have defined a resource script which is to start the Database and listener service during failover. The script is getting run every minute and trying to start the database again and again even though the database has been started already. Is this a general behavior of the Linux cluster? What do I need to perform so that my script is invoked only once? Any assistance will be highly appreciated.

Thanks in Advance!
Khan.

I would also like to add that, the defined cluster service itself is being invoked every few seconds. I felt service should be invoked only during fail over.

In the below status, it’s good to see that once node 1 is brought down the cluster service is shifted to node 2. Well and good here!

Cluster Status for rhel-cluster @ Sun Dec 9 07:54:46 2012
Member Status: Quorate

Member Name ID Status
------ ---- ---- ------
oeltest1.arowana.com 1 Offline
oeltest2.arowana.com 2 Online, Local, rgmanager

Service Name Owner (Last) State
------- ---- ----- ------ -----
service:cluster oeltest2.arowana.com started


In the below status, we see that the node 1 is back up (online) and yet the cluster service is running on node 2. This is an expected behavior!

Cluster Status for rhel-cluster @ Sun Dec 9 07:55:43 2012
Member Status: Quorate

Member Name ID Status
------ ---- ---- ------
oeltest1.arowana.com 1 Online
oeltest2.arowana.com 2 Online, Local, rgmanager

Service Name Owner (Last) State
------- ---- ----- ------ -----
service:cluster oeltest2.arowana.com started


But when you see the below status after sometime, the cluster service is moved back to node 1 which shouldn’t have happened.


Cluster Status for rhel-cluster @ Sun Dec 9 07:57:24 2012
Member Status: Quorate

Member Name ID Status
------ ---- ---- ------
oeltest1.arowana.com 1 Online, rgmanager
oeltest2.arowana.com 2 Online, Local, rgmanager

Service Name Owner (Last) State
------- ---- ----- ------ -----
service:cluster oeltest1.arowana.com started

My question here is why is the cluster service running every now and then when it is supposed to be run only during fail over. This is the reason why my script is being invoked every 30 seconds.

[khan_dba]

Any one with any suggestion?