I'm trying test of failover / failback of NSD. When I removed LUN from a GPFS node, the GPFS node detected the SAN failure and failover to network access via other NSD server.
> Thu Dec 13 12:16:49.065 2012: Local access to nsd2 failed with EIO, switching to access the disk remotely.
Then I recovered the LUN to the GPFS node. According to errpt, the LUN was detected and became available for AIX. In this case, is failback of the LUN (NSD) automatically performed by GPFS or need to run "mmnsddiscover -d xxxx" manually ? I observed that the failback was automatic in most case, but it didn't recovered sometimes and need to run "mmnsddiscover -d xxxx".
> Thu Dec 13 12:19:54.032 2012: Rediscovered local access to nsd2.
I'd like to understand which is the expected behaviour of GPFS.
Thank you for your help,
Pinned topic Is Failback of LUN (NSD) automatic or manual ?
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
Updated on 2013-02-18T15:40:46Z at 2013-02-18T15:40:46Z by ezhong
ezhong 060001JWNJ33 Posts
Re: Is Failback of LUN (NSD) automatic or manual ?2013-02-18T15:40:46ZThis is the accepted answer. This is the accepted answer.>I observed that the failback was automatic in most case, but it didn't recovered sometimes and need to run "mmnsddiscover -d xxxx".
When you read from or write to the lun again, GPFS code will try to rediscover local access to the lun.