Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

Showing and monitoring SCSI-3 reservations in VOM

Created: 18 Apr 2011 • Updated: 29 Apr 2011 | 1 comment
bsobek's picture
0 Agree
0 Disagree
0 0 Votes
Login to vote
Status: In Review

2 Datacenters – each has a Symmetrix and a Solaris-based cluster pair (storage is mirrored by SRDF)

- Datacenter 2 is the active site (SCSI-3 reservations on disk)
- Datacenter 2 failure due to power-outage or other problems
- Failover to datacenter 1 – but SCSI-3 reservations in datacenter 2 remain (but not easily visible --> that’s where VOM could help!)
- There is no failback done to DC2
- After a while the cluster-hw gets upgraded – no issues in datacenter 1
In datacenter 2 we realize, that we cannot access the disks anymore, as the SCSI-3 reservations are on. Nobody realizes at that time why. We were not sure, if we can simply delete the reservations – and even once we tried to, we needed a EMC created script to do so.

It might be very helpful to show the relationship between SCSI-3 reservations and clusternodes – and potentially orphaned reservations (esspecially on RDF2 LUNs)
In the above case, it might have save a lot of time – and I think this is not a singular case...

Comments 1 CommentJump to latest comment

CMilani's picture

We have the CoordPoint agent in VCS that monitors the registrations on coordinator disks. In addition, starting with VCS 5.1, we also have the cluster ID encoded in the registration we place on the coordinator disks. So it’s easy to identify which cluster placed the registration.
 
We are planning on encoding the cluster ID in reservation on data disks via VCS. Once the solution is implemented in the VCS agent, we should expose it in VOM

 

0
Login to vote