VERITAS Cluster Server (tm) startup procedure after power outage or similar events

Article:TECH83007  |  Created: 2003-01-15  |  Updated: 2008-01-08  |  Article URL http://www.symantec.com/docs/TECH83007
Article Type
Technical Solution

Product(s)

Environment

Issue



VERITAS Cluster Server (tm) startup procedure after power outage or similar events

Solution



Proper procedure of restarting servers after a power failure or similar event will ensure that Cluster Server will restart successfully. This example assumes that entire SAN and servers are powered off:

1. Restart SAN and its components (storage, switches, etc) as per manufacturer's recommendation
2. Restart one node of the Cluster Server cluster. Wait one minute *
3. Restart the next node of the cluster. Wait 30 seconds *
4. Repeat steps 2-4 until all nodes in the cluster are restarted
5. Wait for 30 seconds after the last node finishes OS startup *
6. Open a command prompt on one of the nodes
7. Issue a command: hasys -state
8. All nodes should be running
If Cluster Server nodes  display status as STALE_ADMIN_WAIT,  refer to the related TechNote below.
* The intervals cited are not exact. These represent time between restarts that will probably keep more than one node from starting Cluster Server at the same time. The goal is to stagger Cluster Server startups so that the result is not a STALE_ADMIN_WAIT state. The STALE_ADMIN_WAIT state is the result of Cluster Server not knowing which one of the server's differing configurations to use.



Legacy ID



261694


Article URL http://www.symantec.com/docs/TECH83007


Terms of use for this information are found in Legal Notices