Seagate Ultra320's in a RAID array 2

[wbodger]

Now, I have 'inherited' this box, so not sure if it is running RAID 1 or 5. My guess is 5 as it is 4 drives in an array for the OS and SQL Server on our production box. Two questions:
1) How do I determine which RAID was used when this was setup (it has an Adaptec 2000s controller), do I have to re-boot and go into the RAID config utility?
2) If RAID 5 and drive 1 fails, can I take that drive out and move drive 4 into that slot, then put another backup into slot 4 once it arrives?

Or have I totally confused how a proper RAID array is supposed to be setup?

Willie

 

[technome]

Unless you have a utility running from within an OS which can access the raid card you will need to reboot and go into the raid bios to see the configuration.

If you do have a hotspare, which you can find out once in the raid bios, if disk 1 is "pulled" "offlines" or "fails", the hotspare should take over for the failed drive, and the array should rebuild.

If drive 1 (or any other disk) is pulled, fails etc you do not want to remove ANY other disk from the array to replace it, including the hot spare, or you WILL destroy the array and likely all your data, with no other option but to proceed raid recovery, likely costing thousands.

Again, if the hot spare is properly setup it should take over for a failed drive, but it must be designated as a hot spare to do so.

Since your a newbie, shut down the server, pull each disk out individually and mark the disks as to which slot they come from, place them back in. I also mark the cable connectors, all of which keeps you from killing your array in a panic situation, especially the ones at 4 AM.


........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[wahnula]

Good stuff technome...I would personally wait until the server could be restarted before I did ANYTHING...there will be a key combo prompt during POST, and usually after the BIOS IDs the RAID controller. Press whatever combo it requests and get into the RAID BIOS, there you will see exactly what you've got...RAID 1, 5 or even maybe 1+0. I would NOT start yanking disks without knowing the full makeup of the array (if it IS an array).

While I agree with marking all the drives and cables, I would do it with the server powered down. 3Ware controllers don't care which disk goes where, but everyone's different. Caution should be your watchword during the process of ID'ing your array.

Tony

Users helping Users...

 

[wbodger]

Good points, thank you. I know the box has an Adaptec 2000s RAID Array card and it was *supposed* to have come pre-configured as a RAID 5 array, 3 73 Gb Drives and one hot swappable. Now, I have to figure out if it actually *was* configured that way and then if I have a dead HDD or a dying RAID Card...

Willie

 

[wahnula]

Well, if you KNOW that the drives are 73 GB then the array's total size should be 146 GB for RAID 5 plus hot spare. It would also be 146 GB for RAID 1+0, with no hot spare. RAID 1 requires exactly (2) hard drives, so that's doubtful. JBOD or RAID 0 would show 292 GB total.

Tony

Users helping Users...

 

[wbodger]

Yup, that is the size that I have. The apparently failing drive was still under warranty, so Seagate is sending a replacement drive. I'll place that in there and see where we are from there. Thanks for the help and info folks.

Willie

 

[wahnula]

Thanks for the update Willie good luck going forward.

Tony

Users helping Users...

 

[wbodger]

So, the replacement drive arrived today, however... When I went in to replace the 'bad' drive, it was no longer 'bad'. The red light was off and it was blipping away along with it's compatriots... On another note, would it be reasonable to ask about the Adaptec Storage Manager here? The server is a SuperMicro Superserver 8042 with a primary and an external RAID array. The primary array is what has been causing me consternation. The board has built in SCSI support, so there is just an added Adaptec 2000s RAID card. There are 4 10k RPM SCSI drives in a RAID 5 array (one is a hot spare). So, when I open the Adaptec Storage Manager, it shows the HBA, but when I select to 'Switch the View', there is nothing and if I try to create an array group it says I can't do that, I need to setup communication and the 'None of your HBAs are capable of supporting RAID' and yet I have 2 RAID arrays on that server. I am at a loss and wondering if I really just need to reboot the machine and go into the RAID setup on boot. Thoughts?

Willie

 

[wahnula]

The drive could have had some bad sectors detected by S.M.A.R.T.and been repaired by the controller, I've had that happen before on my 3Ware controller...but that just means there are no problems NOW. A drive that "stuttered" gets replaced in my book, even if it is apparently working OK. There could also have been a "hiccup" in the RAID 5 controller, a meaningless glitch.

Of course you cannot create a new array with your Host Bus Adapter, all your disks are in use! I can't explain the empty Storage Manager GUI, your arrays SHOULD be listed there.

Especially in the case of an inherited server, I would wait for some off-hours, and do a re-boot. Make sure you log everything that you see in BIOS, you might need to reset the machine during POST if things happen too fast.

Tony

Users helping Users...

 

[wbodger]

Sounds good, thanks again for the help, I appreciate it!

 

[technome]

Agree with Wahnula, give odds 7 out of 10 odds the disk will will again be an issue, I give a 2 for 10 for the possibility of a disk firmware/driver issue as the cause, 1 that the disk will live a normal life.
The replacement disk Seagate sent you may be new or a refurb, but odds are it will be more reliable then the disk which flat lined already, I would replace it; there is the possibility the disk Seagate sent is the one I sent back last month <grin>
Possible your raid monitoring software is old or an MS patch or system file is causing the communication error or one of the services needed is not starting.
I would definitely look into log files (software or in bios) as to errors found on the individual disks, and run a few consistency checks on the array after you replace the mis-behaving disk.



........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm