EVA5000 mult disk failure message

[noelski]

Just want to ask something about EVA5000 OCP messages, our client encountered the "System Inoperative mult disk failure" on both controller.After a while, the controllers will reboot, will stuck on Scanning Disk and will yield the same message ( "System Inoperative mult disk failure" ).At Command View EVA, no controllers found and we had no option for proper shutdown.

After doing isolations, we finally had it working again by powering off the whole rack ( controllers,disk enclosure,loop switch and the SMA ).What could have caused the problem?The client normally do a snapclone and deletes the used virtual disk afterwards.What can we do to avoid such cases?

Please kindly assists me on this one as our client is asking for a report.Thank you for the kindest response.

[Rob Leadbeater]

Hi,

HP should anaylse the controller logs to determine exactly what happened.

I'll take a guess that a failed or failing disk was creating so many loop errors that access to all of the disks on a particular shelf was lost, causing the EVA to shutdown...

Cheers,

Rob

[grahamswilson]

I had this once - it was as Rob said...

In fact, with the "bad" disk in my case there was no red failed light either - so only after diagnosing the output directly from the back of the controllers was HP able to pin point what disk was causing the issue.

Once that disk was then removed, everything was then OK.

What VCS version is running?

[Patrick Terlisten]

Hello,

log a call at the HP Support Center for log analysis. Maybe a disk hast blocked loops or anything else.

Best regards,
Patrick

[noelski]

thanks for all the response....vcs version is 3.028.actually I forget to mention that during the isolation, two of the disks turned RED for awhile.We also suspected that it could be causing loop errors..that is why our final recourse is to power off the whole rack.Looking at all the LED indicators, no sign of failure except for those two disks.

thanks to all.

[Víctor Cespón · Accepted solution]

If there's a disk or a cable causing problems on the loops, the controllers will try to insolate it running a loop initializacion several times.
But sometimes this fails, and a disk can cause so many problems on both loops that the controllers start losing communication with several disks and mark them as failed.
This can lead to the disk group to become inoperative.

The way out of there is to look at the logs and try to see which disk is causing the problems, remove it and then unbypass the other disks that are not really failed.

[noelski]

thanks to all