RAID trouble

[Mallomar]

I'm not technical, so please bear with me. I have a custom-built PC (almost 2 yrs old) with an LSI SCSI Megaraid 320-2x. The drives are Seagate 15K 147GB SCA drives.

The PC started beeping in the middle of the night, and by the time I got downstairs to check it out, it was doing a continuous (and loud and obnoxious) beeeeeeeeeeep noise. The top drive in the RAID had a red light (others were green).

The system was still running -- ironically, I was doing a backup to an external drive at the time -- but of course I shut down immediately.

The side of the PC case felt hotter than usual.

I tried to find documention on the LSI website to find out what the beeeeep and red light meant, but couldn't find anything.

So I don't know if the drive is kaput, or if it was just warning me that the box was overheating, or what.

And now I'm afraid to turn the PC back on.

Any suggestions/ideas? Could have been a fan failure warning or something? If the drive is kaput, I know I have software that will supposedly restore all of my data, but can I do that before I replace the drive, or do I need to wait until I get a new drive?

I will be calling the builder later this morning when his shop is open, but in the meantime I'm fretting.

TIA,

mallomar

 

[RadiclDreamer]

The beep could be from an overheating, shutting down was a smart thing to do. You really do ned to turn it back on to see what damage if any has occured. It is likely that a fan failed and the system was warning you that it was overheating. Also, depending on the type of raid, you may just need to replace the drive and start a rebuild.

 

[Mallomar]

Thanks for your response.

I've confirmed that the drive is kaput. I went into the configuration/bios thing which told me which drive had failed. I've ordered a new drive (two, actually -- I figure that since I'm buying an extra, I will never have another drive failure!) which is being overnighted.

I could still use the computer -- the RAID has kept it functional, although it's slower -- but the incessant screeching of the RAID alarm is unbearable. I called LSI and they walked me through the process for disabling the alarm, but it didn't work. It's like sitting next to a smoke alarm. The tech guy was perplexed -- he couldn't figure out why it wasn't working. But he was very sympathetic.

Anyway, my plan is to do a total backup of the computer tonight (if the beeeeeeep doesn't keep me awake) and then do the RAID restore thing after the new drive arrives. The builder told me I'll actually to be able use the computer while the restore process is going on. Amazing!

I just really wish I could disable that RAID alarm. I've got work I should be doing.

 

[RebateMonger]

If you have more than one computer, you could use the other computer and work remotely on the screeching one. Move the screeching computer elsewhere.

I've never played with those alarms, but the noise might be damped by taping something over the "horn", too.

 

[Paperlantern]

What RAID are you running? I assumed 5 since you said "others were green" which would indicate multiple others in addition to the failed drive. If it were me I would keep the computer off until the drives arrive and only turn it on when im ready to install a replacement. If you lose one more drive, that array will be toast, and running it is just asking for trouble. Course thats IF it is in fact RAID 5, you could be running some other configuration.

 

[Mallomar]

Paperlantern,

Yes, according to my notes (from when I first bought the PC), it's a RAID 5. I chose it because I have a black thumb with hard drives (my husband says it's because my filenames are too long, but I think that's a silly theory). I've always bought what I think are good-quality drives (IBM Ultrastars, Seagate Cheetahs), but sooner or later I get failures.

My big mistake was not buying an extra drive (hot spare?) when I bought the PC. I thought about it, but the hubster talked me out of it. If I'd had an extra drive, I'd be back at work now. But I'm happy that this all happened on a weekday, and I was able to find a supplier who had the drives in stock and would overnight ship them to me. I realized after I ordered two new drives that one would have been enough -- I'll be returning the dead one to Seagate for replacement since it's still well within warranty. But what the heck. The extra new drive will be insurance.

That's a good point about not using the computer (before I install the new drive) and risking losing another drive. I think what I'll do is turn on the computer briefly and copy my most recent files onto an external drive.

I do have recent backups (I alternate backing up to two different externals, in case one of those fails), and I already copied the most important recent files, so I should be in pretty good shape. Just a few more recent files I want to snag, just in case.

I was thinking that a complete system backup would be good to have in case of a total meltdown tomorrow. But I agree that the risk is too high.

What's your feeling about my computer guy telling me I can actually use the computer while it's rebuilding/restoring (don't know the correct term) the RAID after I install the new drive tomorrow? I was told it would take 8-10 hours to rebuild the RAID.

I have a big project due first thing Monday morning -- deadline is set in concrete -- so it would be great if I could get started on it ASAP. But if it's safer to wait until the RAID is fully rebuilt, I'll wait.

I assume the PC needs to be off when I install the new drive, right? It's got some kind of battery backup thing that keeps fans going even when it's shut down, so I'll have to unplug it. Or do I install the new drive while the PC is running??

 

[Mallomar]

RebateMonger,

That's an intriguing idea, but I have no idea how to implement it. I do have all of the computers networked, but only to share drives or folders, printers, and the broadband connection. I don't know how to actually share applications.

Anyway, I've decided to be cautious and avoid using the PC until I get the new drive installed. Too nervous about possibly losing a second drive, which would be a disaster.

But I would like to know how to work remotely, for future reference -- that could really come in handy even when I'm not dealing with a screeching computer.

 

[Paperlantern]

If both machines are XP/Vista/Win7, you can remote desktop to the one thats screaming from another location to keep from being bothered by the screaming sound.

Yes is it safe to use the machine while the Array is rebuilding, it will just make it take a little longer. RAID is a VERY common technology used for redundancy and QoS for servers. Uptime is crucial for most server implementations of RAID and 99% of all RAID arrays are made to be hot swappable. Data needs to be made available to users while the array is being rebuilt. The alarm is similar to any other mission critical device, like a UPS, most UPS's will start screaming when they are no longer being supplied with electricity to let someone, anyone, know that you are now on battery and have limited operating time. Ideally, if you had the spare on hand, the alarm would sound, you would swap the drive, and the alarm would silence once the new drive was in place and rebuilding. Its not meant to be operated in an alarm state for more than a few minutes when someone is within earshot.


EDIT* Heh, 1024th post!

 

[Mallomar]

Paperlantern,

I'm running XP, but I don't know how to remote desktop.

Anyway, my new drives arrived right on schedule last Saturday, and I decided not to use the computer while the array was rebuilding. I didn't want to slow it down. It took six hours to rebuild (which was less time that I expected), and now the system seems to be back to normal. And the computer's running nice and cool again.

I know the RAID alarm is necessary, but LSI's procedure for temporarily disabling it did not work. The tech guy was perplexed. There were two modes -- disable until a reboot, or disable until re-enabled. We tried both, and neither worked. (My UPS does scream at me when we have power failures, but at least I can shut that off with the push of a button.)

But I learned a lesson from this. I'll have two spare drives -- the other new one I got last Saturday, and the drive (refurb, I'm sure) that Seagate will send to replace the dead one.

All in all, I feel pretty lucky -- the drive failed on a weekday and I was able to find the same drive in stock at a vendor who was able to ship for Saturday delivery. Could have been much worse.