MSA30 Failed Disk

[peakfreak]

Hi All
I have inherited an MSA30 hanging off the back of my SQL Server which is set up in the following config:-

Slots 1 + 2 = 2 x 72Gb RAID 1 Operating System
Slot 3 = 72Gb Windows Page File/Temp.
Slots 4-7 = Empty
Slots 8-11 = 4 x 72gb RAID 10 DB Store.
Slot 12 = 72Gb Hot Spare.
Slots 13-14 = Empty.

Due to a power issue to the building the disk in slot 2 is now showing the red light of doom (Failed Disk). The server is booting and running but it will be in a degraded state I am speculating.

I have 2 questions:

1. Should the hot spare automatically take over the role of the failed disk or do I need to do something to bring it online, if so what do I need to do?

2. Once I get the replacement drive, I believe the MSA 30 is hot swappable, so do I litterally just pull out disk 2 without shutting down the sever, replace the disk in the caddy and then slot the caddy back in place and data/RAID 1 will build automatically.

[Rob Leadbeater]

Hi,

The MSA30 provides no RAID functionality itself - it is just a disk shelf.

Any steps required to replace failed disks will depend on the type of RAID controller that the MSA30 is connected to.

I'll guess it's probably a Windows box, and that the server has SmartArray controller of some form. You should hopefully have some utilities on the server that will allow you to interrogate the state of the arrays.

Cheers,

Rob

[peakfreak]

Hi Rob

Thanks for your reply. I understand that the RAID is handled by the controller installed in the server.
It is a Windows 2k3 box and I have run the HP ADU and ACU and they both report the failed drive as either bad or missing and to replace the drive. I have ordered the new drive but as mentioned in my OP, I am a little nervous about the actual replacement.
I am speculating I have one of 2 options:

1. Shutdown the server throw in the replacement drive and hopefully the RAID controller will automatically rebuild the data.

2. Don't shutdown the server throw in the replacement drive and hopefully the RAID controller will automatically rebuild the data and the server doesn't barf.

I've been very lucky in the past and never had a HDD fail, unfortunately it leaves my experience of Hot/Warm/Cold swapping drives in RAID a little lacking.

[Paul Jerrom]

Howdy,

I BELIEVE (ho ho) that you declare a hot spare for a particular RAID array, rather than for all arrays. My guess would be that the hot spare is set up for the RAID-10 array. If that is the case then you should, using ACU, be able to remove the hot spare as a standby for the RAID-10 array and define it as a spare for the RAID-1 array.
At that point it should (ho ho again) start rebuilding automagically with no need to reboot.
Then when the replacement arrives declare it as the RAID-10 hot spare again.
But much of this would depend on the RAID controller.

Have fun,

PJ

[Rob Leadbeater]

Hi,

If you've run ACU and ADU then you have a SmartArray controller of some form.

Swap the drive with without shutting down the server.

If you pull the drive when the server is down, the controller can get confused as to what's happened with the drive.

Cheers,

Rob

[gregersenj]

Just want to add in.

Assuming its a smart array.

The hotspare do take over automatically. All you have to do, is to call HP and have them send a new disk.

When you replace the defective drive.
Pull out the bad drive.
Wait at least 30 seconds.
Insert the new drive.
In a few seconds it will start blinking.
The Smart array will automatically rebuild to the new disk, and the hotspare wil become hotspare again (The hotspare is shut down when it is not in use.)

Note:
It is possible to assign the same drive as hotspare for all Arrays.
Ofcourse only 1 array can use it. So in case of mulitble HDD failures - 1st. failed drive get the spare.

Note 2:
I have seen SAS or SATA disk, the new get red light when they are insertet.
Just waite a few minutes, it will come on-line.

The swap file might not be importent.
But why not buy an extra disk, and put it on a RAID 1. Can be done on-line.

BR
/jag