In some bureau, a two-node cluster VCS with SQL database was used. When the clients shut down both servers, and turn on one server, it was found that the shared disk can’t be mounted and JAVA console managing VCS can’t be used either. But when another server was powered on and entered windows, the server of master node can mount the disk automatically and the application service of the related database can start up normally. At this moment, the VCS became normal.
Suppose two servers, SERVER1 and SERVER2, while SERVER1 is the master node and SERVER2 is the backup node. Take the following actions:
1. First, run “net stop had” in command line //stop veritas high availability engine service
2. Run “hastart - onenode” after step 1 has been run successfully //start up the related service of VCS to make the two-node cluster run in fault with single node. At this moment, JAVA console is available.
3. Run the following two commands, where “sqlcluster” is the name of cluster servers of SQL
hagrp -autoenable sqlcluster - sys SERVER1
hagrp -autoenable sqlcluster - sys SERVER2
4. Enter the JAVA console, click right on “Source Group”->“online”->“SERVER1”(suppose it is SERVER1 that is turned on)
It is normal in the two-node cluster VCS. Check the documents and it is said that the VCS was running in fault. It can’t mount storage automatically and provide related application services until the fault is fixed.
If you had the problem that both servers of two-node cluster VCS are shut down and can’t be turned on in the meanwhile, and you want the device to provide services, take the actions mentioned above.