An error is received when trying to log in to a cluster through VERITAS Cluster Server Cluster Manager.

Article:TECH23672  |  Created: 2003-01-13  |  Updated: 2008-01-20  |  Article URL http://www.symantec.com/docs/TECH23672
Article Type
Technical Solution

Product(s)

Environment

Issue



An error is received when trying to log in to a cluster through VERITAS Cluster Server Cluster Manager.

Error



Server has failed to respond. Try again later

Solution



This error (Figure 1) occurs if a cluster is in a STALE_ADMIN_WAIT condition.  STALE_ADMIN_WAIT (Figure 2) indicates the system has a stale configuration and there is no other system in a state of RUNNING from which to retrieve a configuration.
Figure 1
 
Figure 2
 

Possible reasons for this include:
1. Stopping the High Availability Daemon (HAD) service on all nodes without closing the configuration (i.e. using hastop -all -force), or rebooting all cluster nodes at the same time.
2. An invalid configuration file on all nodes.
If the administrator is confident that a valid configuration exists on a particular node, that node can be "forced" to read its own configuration file. It is recommended that the existing configurations on all nodes be backed up using the hagetcf -conf command.
Once a backup exists, execute the following command: hasys -force system_name (Figure 3).
Figure 3
 
Node name refers to the system which has the valid configuration file. Ensure that the same case sensitivity is used for the node name. Use the node name case sensitivity that is in Figure 2. This command will force the node specified to read its local configuration file. If the configuration file is valid, the node should transition from LOCAL_BUILD to RUNNING. All other nodes would then show as REMOTE_BUILD and transition to RUNNING.



Legacy ID



253696


Article URL http://www.symantec.com/docs/TECH23672


Terms of use for this information are found in Legal Notices