3ware card reports SECTOR REPAIR. Should I start worrying?

Posted by: Roger

3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 07:46

Windows 2003 x64, 3ware 8504LP, 4x200Gb in RAID5. I've seen the following in my event log:

Code:
Mar 11 09:38  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 10 20:00  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 10 16:44  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 09 22:52  FwReset> 2 outstanding I/O's
Mar 09 22:52  The driver for device \Device\Scsi\3wDrv1001 detected a port timeout
              due to prolonged inactivity.
              All associated busses were reset in an effort to clear the condition.


The earliest timestamp is Mar 09, because I only installed Windows 2003 on this box on Friday. Before that, it was running Ubuntu.

I'm currently restoring my FLAC collection from backup and listening to some MP3s on a different computer, so the disk's getting hammered reasonably hard. Oh, and the music skipped at the time of the last one.

Are these warnings the sign of an impending disk failure? Should I replace the disk on port 3? I've got a couple of spare 200Gb disks banging around here somewhere. Or should I wait until these kinds of things are happening more frequently, and when the computer's not under load (i.e. closer to normal use patterns)?
Posted by: lectric

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 07:56

How critical is the data on the drives? Personally, if a drive kicks out once, it was weird. If it happens again soon, it gets replaced. Soon is a relative term depending on the criticality of the data. If it's a production server at work, a drive getting dropped twice in six months is too often. At home, if it happens twice in a week, I may just backup more carefully for a while.

-=edit=-
And from what I've seen, that's too often for even at home. If it were me, I'd sleep much better if I had replaced it. It may never fail, but how much is your time worth if you had to restore from backups?
Posted by: Roger

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 08:14

Quote:
If it were me, I'd sleep much better if I had replaced it. It may never fail, but how much is your time worth if you had to restore from backups?


Yeah, this is at home, and I have recent backups.

Thanks for the advice. What I'll do then is this:

1. Wait for the restore that's currently under way to finish -- because I've moved the box from Linux to Windows, I'm having to use another Linux box to read the files from my (ext2) backup media and squirt that over the network, so I'd like to finish the restore first.
2. Take a fresh backup, in a Windows-compatible format. I'll probably just use NTBACKUP for this until I find something better.
3. Swap the disk on port 3 for another one.
4. Rebuild the array.
Posted by: mlord

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 13:51

Quote:
Windows 2003 x64, 3ware 8504LP, 4x200Gb in RAID5. I've seen the following in my event log:

Code:
Mar 11 09:38  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 10 20:00  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 10 16:44  AenAddQ> Port 3: SECTOR REPAIR (0x23)
Mar 09 22:52  FwReset> 2 outstanding I/O's
Mar 09 22:52  The driver for device \Device\Scsi\3wDrv1001 detected a port timeout
              due to prolonged inactivity.
              All associated busses were reset in an effort to clear the condition.




Because this is closed source software, you'll have no idea what it really means unless they've documented it somewhere.

Or unless you pull the drive from Port3, connect it to a Linux box, and do "smartctl -a" on it to read the error logs.

This could be totally harmless (ECC report --> found/fixed a bit error, happens all of the time). Or potentially not so good (bad sector, got remapped when overwritten).

Cheers
Posted by: mlord

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 13:54

Hmmm.. based on reading this, it sounds like a genuinely bad sector was found and fixed by the array management software. If it is a genuine RAID (not RAID0), then no data was lost.

Cheers
Posted by: Roger

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 15:36

Quote:
If it is a genuine RAID (not RAID0), then no data was lost.


It's definitely RAID5, so I'm not going to worry unduly. I'll schedule a verify pass to see if anything else is wrong.
Posted by: oliver

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 11/03/2007 22:20

Quote:
installed Windows 2003 on this box on Friday. Before that, it was running Ubuntu.


I'm really curious why you made that switch, because I've been in the process of going the other way.
Posted by: Roger

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 12/03/2007 07:50

Quote:
I'm really curious why you made that switch, because I've been in the process of going the other way.


Mainly for the hell of it. I had a licensed copy of Windows 2003 hanging around, so I thought I'd get some use out of it, and some experience with it.

I've not completely abandoned Linux -- this is just my fileserver (where the OS doesn't particularly matter from a technical standpoint). I'm still running Ubuntu (6.10 Desktop) on my laptop and Debian on my web server.

I'm also running Ubuntu (6.10 Server) on a virtual machine (under VMware) on the Windows 2003 machine, in order to handle those things (remotely backing up the web server, for example) that work better on Linux.

Given that I work at a company that does Windows infrastructure management (software and consultancy), getting more experience with the things we're supposed to support is a good thing, career-wise.

If this isn't your situation, then I'd encourage you to try Linux (my current favoured distro is Ubuntu), to see if it fits your needs.
Posted by: Roger

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 12/03/2007 07:52

Quote:
I'll schedule a verify pass to see if anything else is wrong.


That aborted with "Verify Failed: Unknown error". Windows reports something along the lines of "Disk on port 3 failed. Array degraded."

Nice. Also turns out that none of the spare disks I had are SATA, so that means a shopping trip today.
Posted by: Zsolt

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 27/01/2013 00:06

First check for vibration. If the PC house is vibrating change your ventillators. Had lot of trouble until (finally) found out this simple solution.

No sector repairs (and other fancy things since).

Zsolt
Posted by: JBjorgen

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 27/01/2013 01:49

necropost
Posted by: Shonky

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 28/01/2013 15:29

Future spammer?
Posted by: gbeer

Re: 3ware card reports SECTOR REPAIR. Should I start worrying? - 29/01/2013 00:41

Googlodyte?