I'm running sun cluster 3.1 and SVM 11.9 and Solaris 9 (5.9 Generic_122300-57 sun4u sparc SUNW,Sun-Fire)
After syncing mirrors from old 9970 and new Hitachi storage and detaching old 9970 submirrors, we are trying to remove old 9970 did device from diskset.
This is what i got:
root@node1 # metaset -s diag-set -d -f /dev/did/rdsk/d18
metaset: node1: diag-set: host node2 does not have set
I can switch the service-group and also devices from node1 to node2 and vice versa.
I've already did the same steps over another cluster hosted by the same 6900 without erros.
Thank's in advance for replies.
quite old configuration but maybe this doc will help:
I didn't try that procedure yet, hosts are in production enviroment and i need to evaluate any possible impact with the customer.
What about clustered services such as Oracle (failover) instance configured on that diskset (diag-set)? Do i need to disable the service or i can try while Oracle is Online on the other node?.
The steps only make changes to the metaset, it removes and then adds the host that gives the error.
There is no impact on your cluster. The worst that can happen is that you may see errors from the metaset
if the issue is not resolved after these procedure.
You could minimize the impact by disabling all the resources in that resource group including the HAStoragePlus
that has mounts associated with diag-set. Then run the procedure and see if you can release the set and take
it on node2 without errors. If it worked then you can enable all resources and test a resourcegroup switch.