Storage System Not Recognized

[Richard W. Schierer]

Background...
Compaq Proliant 1600, WIN NT4.0, 1GB memory, 3-4.3GB hard drives, 2-external disk arrays; Compaq Storage Systems F1 with 5-18.3GB hot swappable SCSI drives in each. It is running SQL Server 2000 and using Slot 2's F1 as the 'live' database [this is a Compaq Smart Array 221 Controller]. Slot 3's controller is being used to control the 1600's internal SCSI's which are the OS and boot, and the backup F1 [this is a Compaq Smart-2DH Array Controller. During the evening, SQL Server backups up the live data from Slot 2's F1 to Slot 3's F1. Then at Midnight, Slot 3's F1 is backup using Arcserve on another server and tape drive.

Problem...
One of the drives on Slot 3's F1 displayed an amber light indicating it had failed. I bought 2 new replacements, which were not exact, as they had discontinued that model. The new drives are slightly bigger. The second one is going to be used for a spare, in case of future problems. I have discovered that I made a
'BIG'mistake, when I powered off SLOT 3's F1, without first shutting down the Proliant 1600. I found a tech note that said that by doing this, the system had now marked the other 4 drives as failed. When I put in the new drive and powered it up first with the 1600 off, only the new drive's lights flashed. The other 4 showed no indicator lights during the entire boot process. Only the new one flashed a light interminiently during the boot process.

When the server was up, I went into the Compaq Array Configuration Utility. When I selected the SLOT 3 array, it showed ARRAY A with the 3-4.3GB drives with an A with a red diamond around it signifying to me that these were already assigned to an array. It also showed me 4 drives that were bad and one good drive. I contacted Compaq Tech support via email and explained the prolbem. They informed me to delete the array and reboot, which I did. When I went back into the Array Configuration Utility, it showed only the new drive that I had installed and not the other 4. When I tried to create a new array, it again only showed the one drive. I checked this in logical drive and physical drive views.

I have shutdown both machines [1600 and Slot 3 F1] and removed the power cables and controller cables and reseated them. I powered up first the Slot 3 F1 and waited for several minutes and then powered on the 1600. When the Slot 3 F1 came on, [and this is only after several hours and reboots AND removing each drive and reinserting] all the lights flashed on the Slot 3 F1 when I turned it on. When the 1600 was powered on, it's array's lights came on, but the Slot 3 F1 showed no activity at all. During the bootup process, the 1600 displayed the following...
Slot3 Compaq Smart-2DH Array Controller, 1775-Slot 3 Drive Array-Proliant Storage Enclosure Not Responding SCSI port2:Check Storage enclosure pwer switch and cables. Turn System power off while checking power and SCSI cable connections. then turn system power back on to retry.

OK, done that all. I tried to shut down just the Slot 3 F1 and then power it back on with the 1600 on. I shut them both off and then the F1 on first and then the 1600 on. I still get the same error message and when I go into the configuration utility, it still only shows one drive available.

I know there must be a way to tell the system that the other 4 drives are there. The other 4 drives can't have failed as well could they? Or the F1 couldn't have failed itself? The other F1 in Slot 2 is working fine and is supporting the live SQL Server 2000 database. I have not shut that one off at all during my troubleshooting. My users are still working with no problems. I haven't had a good backup for almost 2 weeks and I am to say the least 'edgy'.

I like a challenge like anyone else, but this is now past a challenge.

All replies thankfully acknowledged in advanced.


Rich

[Marino Meloni_1 · Accepted solution]

hi,
strange problem, and as I didn't know well the proliant array controller, check if this could apply to your case:
I understand that you had a raid 5 with 5 disk 18gb each.
the steps you performed are the following:
one of this disk fails.
you sutdown the array
you replaced the disk and and a spare
you reboot all and only the nesw drives are seen (you lost the four old dirves)

I suggest you to remove the new two dirves, and reboot with the old ones only(four disk, probably if the system could not see them before,and now it can, they may still have the configuration on it)
if they appear, ask HP-CPQ support how to reload the config form disks, porbably it can be done and then you will recover your raid 5 even if in degraded mode.

If the four disk did not appear, it may be caused by a hardwaree problem on the extermal disk array

[Richard W. Schierer]

The 2nd drive is a spare that was not installed in the array. I just bought it so that if there was a future problem with either one of the F1's that I would have one on hand.

I will definitely try out your suggestion tomorrow morning and let you know how I made out.

When I put it back the way it was, should I put back in the original failed drive? I guess I would have too put the failed one back in so as to have all bays with drives in it.

When I originally replaced the drive, I was getting error 1724 and 1727 messages. These say tht Physical drive position changes were detected and the logical drive configuration has automatically been updated and new logical drive attachment has been detected, Loical drive confiuration has automatically been updated.

That was when I was able to go into the array configuration utility and see the 4 'failed' drives. It was at that time that I deleted them, per Compaq's suggestion. Then when I rebooted they never showed up again.

Thanks for the advice. If you think of anything further, please do drop me a line.

Again all replies thankfully acknowledged in advanced.

Rich

[Richard W. Schierer]

This morning, I removed the new drive and put the old one back in. I powered down the 1600 and the SLOT 3 F1. I brought the F1 backup and then the 1600. Upon booting up the F1, all lights flashed. I then went into the Array Configuration Utility and all 5 drives were shown and I created my B Array. The first row of lights are on now. I guess I will have to reboot the 1600 so that it recognizes it as an array so that I can start using it.

Thank you so much for the advice, but now I am concerned why the amber light was displayed in the first place indicating that the drive had failed. And why wouldn't the new drive allow the others to be recognized? It is a different drive from the other 4.

Rich

[Richard W. Schierer]

Some more information....

I called Maxtor tech support and asked them if the hard drive that I was using to replace the failed drive was compatible with the other hard drives and the F1. They told me that I had to use 5 of the same drives in the array. So now I am going to have to buy a total of 5 of the new drive so that I can have 5 - 18.3GB drives.

I am now waiting for a response back from Compaq tech support to find out if the Maxtor Quantum Atlas 10K III 18.3GB Wide Fast SCSI drives with work in the Compaq Storage System F1 with the Compaq Smart - 2DH Array Controller.

Rich

[Marino Meloni_1]

hi
for sure the hdd that are part of an array have to be compatible, id didn't mind who is the vendor, but they need to have same caracteristics. Hp qualifie disk from differnets vendors with the same part number, these disks could be exchanged without problems, so ask CPQ support, with the p/n of all disks, if they are compatible with the system, they will have the answer.
Also, if one of the disks have some kind of incompatibility or have some trouble on the scsi interface, it could "hang" the scsi bus and prevent other disk on the same but to work and to be seen by HBA.
marino

[Frank Kaufman]

Actually the size of the drive does not matter as long as it is the same size or larger then the other drives in the array. You will lose the extra as the SMART Array controller will only use the same amount as the other drives. We regularly replace failed 4.3 and 9.1GB drives with 18.2-s, as they are available. We lose the extra capacity but gain a functioning array. I only have experience with Compaq drives though.