DL320 G6 - B110i RAID Controller - Rebuild Raid 1+0

[anyCOPY]

Hi all,

we have a HP DL320 G6 Server, running a Smart Array B110i RAID controller, with 2 x 750GB SATA disks.

Yesterday, one of the disks failed. We have replaced the faulty disk, and in the Option Rom Config ( BIOS ), it states RECOVERING. This message has been standing there now for almost 24 hours. Is this normal behaviour? Does a RAID rebuild really take this long?

Attached screenshot of the config screen ( through ILO )

Thanks in advance for your responses.

Kay Dieckmann

anyCOPY ApS

København - Danmark

[-Cisco-]

Hello the issue you are having is not common, can you check the status from the ACU (Array configuration utility)? and maybe post a screenshot.

[anyCOPY]

Hi,

sadly no, we are running ESXi on this server, so no Windows is available. Also, upon booting, I tries to boot up from C:, but stalls there.

Also, the server is located in a datacenter, so at this time, I dont have physical access to the server, unless I drive there. ( which is why I'm on ILO )

Any other thoughts?

[Goutham_Sabala]

Hi, 

One Gb will take about 15mins, this is theorteical value , it completely depends on the controller, rebuilding of the array happens when the controller is free, usually controller that has a cache works faster than the one without, B110 is a dynamic controller which simply means that the load is on the OS for rebuilding, It also depends upon the rebuild prioirty has been setup in the array config utility ( which cannot be changed untill the array is completely up )

[anyCOPY]

Hi Goutham,

Thanks for also answering in my thread.

By your statement, you are contradicting what Cisco said. He claims its not normal behaviour for a RAID rebuild to take this long, while you state that 1GB =15 minutes. With a 750GB SATA drive, that would equal 11.250 minutes=187,5 hours which is roughly 8 days.. for a RAID 1+0 rebuild. I hope you're not serious, because that is unacceptable in any enterprise environment

Also, we're not using an Array Configuration Utility. The screenshot I posted is the only way of telling what the RAID is doing, since we're not running a 'normal' OS on the server, but ESXi ( for 2 virtual machines in VMWare ).

Thanks for your input though, any other thoughts are greatly appreciated

[Torsten.]

>> One Gb will take about 15mins

Based on this it would take a week?

I have seen several cases with a smartarray 410 without cache when recovering took 1 day for a 300GB disk - seems to be normal. Obviously it is much faster with cache.

[anyCOPY]

hi Torsten,

Also thanks for contributing.

Based on what you said, 300GB in 24 hours, I should be able to have my RAID rebuilt tomorrow by the end of the day.

I will show some more patience, and report back in here when/if its done.

Kay Dieckmann

anyCOPY

[Torsten.]

This is a background operation, so you can continue to work. Depending on the load it could take a bit longer to sync the data.

[anyCOPY]

Well, thats the issue, its not a background operation for us.

It was the first disk in the array that failed. We shut down the server, replaced the faulty drive with a spare ( in slot 1 ), and booted the server up. In the Option Rom Configuration for Arrays ( see first post screenshot ) it states that the Status is Recovering, with no progress indicator. If I let the server go through the POST boot, it comes up with a message that it's trying to boot from C: , but nothing happens. So we're stuck with a non-working environment, I guess, untill the rebuild is completed.

In hindsight, should we have placed disk 2 ( the working one ) in slot 1, and put the disk to rebuild to in slot 2?

[Rashwanov]

@anyCOPY Am in the exact same situation now.

Can you please advise what you have done, if you managed to get ESXi boot once more ?

[maabo]

hello, I was facing same issue too, and I discovered, that we were probably dealing with a bug in the firmware. There was unexpected behavior in my case Namelly, there is a differece if the procedure is started by hitting the F1 key or if it is started by countdown to zero

RAID 1 out of sync, B110i, firmware version 1.38, controllers firmware displayed something like:

press F1 for continue with auto data recovery (resync),

press F2 for continue witnout resync 

and it counts down  60 seconds with default option (F1).

but if the F1 button was really pressed, it hangs with no response... I was wandering, if the synchronisation runs at background, so wasted couple of hours waiting for any response from the controller

But I recognised later, that I should wait the 60 seconds for the default action without any KEY PRESSED .... next time, I let the countdown go to 0....

and voila, resync in background was immidiatly started ("automated data recovery in progress" appeared) and system continued with normal boot procedure!

(as with other non HP raid products, the array rebuilds at background with OS normally operating)