Good morning,

I'm running into some problems.

System has been running wonderfully for several years now.

Asus P5B-VM SE, Pentium D 805, 2GB mem

Seasonic S12-430 psu

Promise TX4 PCI controller.

Been running with 4 drives on MB, and 4 drives on the PCI card.

Currently running 5 Beta 2

 

I decided it was time to upgrade.  Picked up an AOC-SAS-MV8.  Started moving my drives over and got errors during a parity check for one disk.  After a lot of trouble, I finally decided to move everything back to the PCI controller.  Ran an Initconfig and rebuilt my parity drive.  Ran a parity check and everything was working fine. 

 

Next, I decided to start from scratch and move everything over to the new controller.  Put everything in assigned my disks.  Initconfig.  Rebuilt parity.  Everything was looking good again until I ran a parity check.  Received errors at the exact same time as before.  576 errors on the same disk at about 12% of the way through.  I've attached part of the syslog.  Any advice would be appreciated.  thanks.

 

 

syslog-2010-11-10.txt

Suggest you post entire syslog.

 

But looks like your problem is with disk6 (ata2).  You might try reseating power and data cables to that drive.  You might also reattach SAS connector to controller card.  My guess is you have some marginal connection issue.

Unfortunately that's all I have saved off the syslog.  Site kept saying it was too large so I trimmed from the front and end.  Guess I should have zipped it but didn't think about it at the time.  But, disk6 definitely seems to be a problem.  As I moved cables back and forth from the different controllers, I definitely reseated and used different cables and produced the exact same error at the same point.

 

Right now I'm copying the data off that drive onto an empty drive in the array just in case it's going bad.  I really don't think it's bad, but might as well be safe.  Then I may just remove the drive from the array and rebuild the parity.  I also just happened to order a new drive the other day so could use that to rebuild.  Got a long time to think about it while the data is copying.

Unfortunately that's all I have saved off the syslog.  Site kept saying it was too large so I trimmed from the front and end.  Guess I should have zipped it but didn't think about it at the time.  But, disk6 definitely seems to be a problem.  As I moved cables back and forth from the different controllers, I definitely reseated and used different cables and produced the exact same error at the same point.

 

Right now I'm copying the data off that drive onto an empty drive in the array just in case it's going bad.  I really don't think it's bad, but might as well be safe.  Then I may just remove the drive from the array and rebuild the parity.  I also just happened to order a new drive the other day so could use that to rebuild.  Got a long time to think about it while the data is copying.

 

The top of the syslog is important.  Cutting out long series of repeating errors is ok.  Sometimes the very bottom of the syslog has some useful info as well.

 

One of the things the syslog would have said is what type of disk is disk6.  What is it? Is it the same model as other disks that are working?

 

Hunting down cabling problems can be a challenge (and normally you don't know if you really have one in the first place!).  Tests I will do ...

 

1 - Swap 2 disks on the same controller (one working, one giving problems) and see if the problem moves to the other disk.  This should involve physically moving the disk into a different backplane drive slot (or, if drives are hardwired, swapping all of the cables at the drive side, not the controller side.)  If the problem moves to the other disk, you know you have some problem in the signal or power path, and that the disk is okay.

 

2 - If the problem follows the disk, swap with a disk on another controller.  If that eliminates the problem, you might just leave it like that.

 

3 - If the problem still follows the disk, you could try hooking the disk directly to the disk controller or motherboard eliminating backplanes and power splitters.  Use a new data cable and a known good power connection.  If the problem still occurs with the disk, it must be the disk.

 

One of the keys to doing this successfully is have a test you can run that reliably recreates the problem, which doesn't seem a problem here.

 

Give these a try and report back.

Here's the current info from the beginning of the syslog with boot and drive info.  It's a 640gb WD.  Different model than all others.  All drives are in an Icy Dock MB455SPF-B.

 

syslog-2010-11-10.txt

Well, I moved the disk to a different slot on the same controller.  Rebuilt the drive successfully.  Ran a parity check and came up with no errors.  I guess I should move it back to the original slot and run a parity check, but I'm tempted to leave it alone for now.  I have another drive on the way which I can test on that slot without messing with my valid array.

Sounds like a good plan. Glad you are back up and running.