Storage Saga So far

[ccrichard]

The other weird things:

1. Failover would not finish

2. The Primary node was showing more than one iscsi connection per lun per client. (Under Maintenance connections). Now it only has one iscsi connection per lun and failover works.

Question:
Is there a way to stop and restart just SCST service and not failover?
Is there a "safe" way to shutdown the primary node in case it does not want to complete failover?

[ccrichard]

I've noticed the vmware/lun locking tends up happen when very disk intensive operations are attempted.

I believe the November problem happened when I was experimenting with an offsite backup job. The backup job succeeded. Very soon after the job finished the logs stop.

The outage in mid December happened when I was trying to move an entire VM through the Vsphere Client.

[ccrichard]

I had also deleted some snapshots recently

[Gr-R]

de-activated from DSS V6?

I suspect a load on the controller, causing the locks.
as things go forward, time will tell.

[ccrichard]

I am suspecting hp's firmware. There are some updates to be applied.
It looks like their smart array controller does not play nicely with sata disks.

The confusing part is that they have two different revision numbers for the firmware update. It could be based on the HW revision of the controller: I'll try searching by S/N and see if it is more specific. if not, I'll do a chat session with HP support.

I'll also see if I can budget for all sas disks and rebuild the array/re-install Open-e.
FYI the snapshots I deleted were vm's, not open-e snapshots.

[ccrichard]

HP has an advisory for the SATA disks and Smart Array controller I was using.

http://h20565.www2.hp.com/portal/sit...4892.492883150

[Gr-R]

http://h20565.www2.hp.com/portal/sit...4892.492883150

Good find!
Seems either a firmware update on the controller, or firmware updates on the drives themselves can solve the HP issue.