Resource state displayed as ALERT

  • 7001043
  • 28-Jul-2008
  • 27-Apr-2012

Environment

Novell NetWare 6.5
Novell Open Enterprise Server 1 (OES 1)
Novell Open Enterprise Server 2 (OES 2)
Novell Open Enterprise Server (Linux based)
Novell Open Enterprise Server (NetWare based)
Novell Cluster Services

Situation

After a node failure the cluster state is listed as FAILBACK ALERT.

e.g.  OES Linux

NODE2:~ # cluster status
Master_IP_Address_Resource         Running    NODE1      5
CPOOL1                             Failback A NODE2      5
CPOOL2                             Running    NODE2      4
DHCP                               Running    NODE3     13
DATA                               Running    NODE4     13

e.g. OES NetWare:

NODE4:~ # cluster status
Master_IP_Address_Resource         Running    NODE1      5
CPOOL1                             Alert      NODE2      5
CPOOL2                             Running    NODE2      4
DHCP                               Running    NODE3     13
DATA                               Running    NODE4     13

Novell Remote Manager (NoRM) and iManager also report the same.

Resolution

Use the following syntax, which will accept or reject an alert on a resource and control manual start, failover, or failback to a node.

   CLUSTER ALERT {resource} {YES / NO}

e.g. using the example, above:

   NODE2:~ # cluster alert cpool1 yes

   Status for Resource: CPOOL1
   Failback Alert on NODE2  Lives: 5
   Revision: 1494023683

This will fail the resource back from to NODE1 from NODE2.

Additional Information

This is working is designed.  A resource will be put in to ALERT state if it is has a Manual Failback policy set in NoRM or iManager.

e.g.  Suppose a resource is on NODE1 and it fails over to NODE2.  When NODE1 is brought back up, the resource will be put in to ALERT state as it requires manual intervention to choose whether to fail back to NODE1 or leave it where it is, on NODE2.  Once a resource is in this state, CLUSTER ALERTmust be used; changing to automatic failback in iManager won't help at this point (but will for future failovers).

For more information refer to the documentation https://www.novell.com/documentation/oes2/clus_admin_lx/data/aj7bqnw.html