DL380 G6 RAID 1+0 Failed.

[Kusado]

Logical Drive consists of 6 drives in 1+0 configuration. One of drives is "Bad or missing". Another one "predicted to fail soon".

Is there any way to save the data on RAID?

As I understand, I should change the Bad drive, wait for rebuild, then change the predicted bad, and the array should go online?

[Suman_1978 · Accepted solution]

Hi,

Theatrically your assumption is correct.

ADU report shows three drives are failed:

Physical Drive 1I:1:2 The physical drive has failed.
Physical Drive 1I:1:4 The physical drive has failed.
Physical Drive 2I:1:7 Physical Drive State: Predictive failure. This physical drive is predicted to fail soon.

At this point, in my view you need to take a full backup before you replace any drive.  The Bad drive may not rebuild because the other drive is marked Predictive failure.

Also in your server the P410i firmware is 2.74, but at HP the new version is 6.40 (B), and ADU version is 8.40 but at HP its 9.40.12.0.

Its recommended that you keep the firmware and software upto date.

Thank You!
I am a HP employee.

[Kusado]

Physical Drive 1I:1:2 is assigned to RAID 1 in another logical drive, so it's not a problem for a while. I just want to be sure that information is restorable before I order new disks.

And what do you mean "full backup"? The logical disk is offline, how can I backup it? Sector copy every disk plugging them to another machine?

[Robert Egloff]

Could he add a 'hot spare' an enable it?

Not sure just what that might do - that's of course, if you have an additional drive.

On the backup - I think the idea was a tape backup or perhaps a backup on another server/USB device.

If I was in your shoes... I'd put in a new drive; mark it as a hot spare and see if it starts a rebuild. 

Otherwise - I'd put in the new drive - configure it as a single drive and dump an image on that new single drive... 

[Kusado]

There is no place to insert hdd for spare.

There are no backups of this exact server as it's a part of HA cluster and there's no need to have a copy of each cluster node. It could be just little simpler to restore existing configuration by taking array online than building it from scratch.

Thanks everyone for your suggestions.