Start a Conversation

Unsolved

This post is more than 5 years old

1903

December 11th, 2015 08:00

Booting problem on an AX4.

We have an AX4 array with a single SP. We had a drive fault on Disk 5 (I believe it is 5). Before we were able to replace this drive, we
had an unscheduled power outage to the array. When power came back, the array started its boot. However, we now have two faulted drives, Disk 5 and Disk 0. I believe Disk 0 is a vault drive.

 

The array has been booting for over 12 hours now (green light on SP still flashing). We can ping the SP, but we cannot run any naviseccli
commands to it (Message: Error occurred because connection refused. Management Server is not running.)

 

How do we determine if it is hung in its boot, or is it rebuilding the vault drive and we just have to wait?

Any thoughts/suggestions?

195 Posts

December 12th, 2015 16:00

One quick thought:

A failed vault drive will not rebuild until the faulted disk has been replaced. 

The vault data can only exist in the designed slots.  If you have a spare anywhere else in the array you could try physically replacing the faulted vault drive with that one.

If you have user data on a vault drive the rebuild of the user data can be...glacial...because there is no write cache available when there is a faulted vault disk.

So it could actually be rebuilding the user data, but you are going to need to replace the failed vault disk before it becomes anything like normal.

Best of luck.

4.5K Posts

December 28th, 2015 14:00

The first 4 disks in the enclosure are the Vault disks (0, 1, 2, 3). Disks 0 and 2 are mirrors of each other - if disk 0 fails, then the array should still be able to boot from disk 2.

You could try removing disk 0 and see if it boots, but if you have configured the vault disks as user storage, that will fault the raid group and if disk 5 is also part of the raid group that uses disk 0, then you would have a double faulted and could lose all data.

If you have a spare that matches disk 0 in size, then you could try replacing disk 0 and reboot.

Otherwise you could contact EMC support.

glen

No Events found!

Top