Why can I not save a syslog to one of the unRAID hard drives?  As you probably have guessed, I'm kind of a novice when it comes to this OS and language.

You can, to /mnt/disk1 for example (cp /var/log/syslog /mnt/disk1/syslog.txt), but I understood that you could not access correctly your flash drive at /boot, which means your networking may not be working right, and the drives would be unreachable across the network.  That's why I suggested a separate flash drive, it's portable.  If the network is good, then yes, copy it to one of them, and transfer it from there.

 

On another note, in regards to the smaller disk3 I removed from my unRAID server and started this whole mess, can I simply put it back as a new drive but not lose the original data, or do I have to use reiserfstool to drag-and-drop all the files (which would take forever)?

 

Everything on that drive should be exactly as it was when you pulled it.  You can re-install it, and once you have re-assigned it (if the Web management page is available), all of its file system will be accessible, and look exactly as you last saw it.

Actually, I can access all of the drives and use the web interface.  Aside from the new drive not being rebuilt with the data of the drive that was pulled, and not being able to properly write a syslog to boot, the unRAID has been working fine (though obviously reporting parity errors). 

 

Thanks for helping me create a syslog on disk1. The zip is 134KB, over the 124KB limit for this forum, and the forum doesn't allow .rar, so here is a link http://www.cyberonic.com/~curtisr/syslog 9-6-08.zip.

I took a look at the syslog and didnt find anything which could cause the mess. There are a lot of parity errors and nothing more. About your incapability to write to the flash. There are some errors like this "Tower kernel: FAT: Directory bread(block xxx) failed" which may mean the stick is broken. Plug it to your PC and scan the surface for errors.

 

You have to wait for one of our linux-guru. Maybe he will find something more in the syslog. In fact i'm a newbie in this matter.

I'll try the stick on the PC & scan it & get back to you.

 

For the heck of it, tried another parity check last night. 12981 errors, which is larger than the last.

I'll try the stick on the PC & scan it & get back to you.

 

For the heck of it, tried another parity check last night. 12981 errors, which is larger than the last.

First thing I would do is run a memory test.  Let it run for several passes, or even overnight.

 

Joe L.

I'll do a mem test later.  Windows scan of the flash drive w/ chkdsk was negative.

Your syslog, especially the bands of parity errors, looks only too similar to the other thread I linked above (here), so it does not seem advisable to try any more to get the current hardware and drive group to work.  I'd remove this new drive and re-install the previous drive, and make sure that everything is stable again.  It will require rebuilding parity.  Then run a parity check to ensure all is well again, should be zero parity errors.

 

As in the other 2 users' syslogs, there are repeating patterns of parity errors, that seem to occur in bands, some of which overlap.  Multiple gigabytes are checked error free, then multiple gigabytes contain errors often every 8 blocks, and then more gigabytes free of errors.  Some of the blocks in error are the same in multiple runs, most of the rest are new each run.  Both types of errors are very wrong, but for very different reasons.  Repeating parity errors are supposed to be impossible under normal conditions.  Two examples, block 283120 occurs in 3 of the 4 parity checks you attempted, and 2354544 occurs in all 4.  It therefore does not look to me as if it is a data corruption issue, at least with the data on the drives.  I suspect some unusual incompatibility with particular combinations of kernel, driver, motherboard, drive controller, and drive, which has proven so far to be hard to pin down.

I'd remove this new drive and re-install the previous drive, and make sure that everything is stable again. 

 

Well, in the process of attaching that removed drive to my PC so I could use some of the data, I accidentally shorted the drive, thus it's dead.  When it rains, it pours.

 

As far as rebuilding the parity, I was thinking also it was a good idea.  Since I have never done that before, I want to make sure I don't screw anything up and lose more data.  Is there a wiki on rebuilding the parity?

Memory test for a couple of hours was without errors.

 

Decided to stop the array & restore, making a new parity from scratch.  When it completed, it reported zero errors. I just copied a very large file to the unRAID and am now doing another parity check.

Guess what? Only @ 3.2% of the parity check & already 14391 errors!

 

I have another, different motherboard & processor laying around if anyone thinks that could be the problem.  I could swap the RAM from a workstation too.

 

Any comments?

Memory test for a couple of hours was without errors.

A good sign

Decided to stop the array & restore, making a new parity from scratch.  When it completed, it reported zero errors. I just copied a very large file to the unRAID and am now doing another parity check.

When doing a "restore" unRAID is building parity for the first time. (or as if it is the first time) 

It is impossible to have a "parity error" when initially writing it. It does not yet exist to compare to a computed value. ... It is possible to have disk "read" errors or disk "write" errors, but not parity errors.

 

I see your subsequent parity compare failed.  You might just want to swap out the motherboard...

 

Joe L.

Okay, here is the latest...

 

I realized that I had added a new drive into my unRAID case, but never plugged in the power (only the SATA cable) or formatted it.  I was hoping this was a funky cause of my problem.  I changed all SATA cables, plugged in the power to this new drive, formatted it, ran new parity.  It gave errors again.  Yesterday I then did another parity restore.  This morning, a new parity check, with many errors  .

 

Next up, changing motherboard, processor, and possibly RAM. Will report back.

Be sure to let us know how it turns out.. 

ZERO PARITY ERRORS!!!

 

Okay, problem fixed (thought I wouldn't say solved).  Last night changed motherboard & processor.  Ran parity check, which of course found thousands of errors.  Ran the subsequent check, and voila, no errors.  Same memory, power supply, cables, case, SATA PCI card.  New board had gigabit LAN, so ditched the PCI one I previously used.

 

What caused this problem is your guess.  Previous board was ECS (have to look up model when I get home) + Intel Duo Core processor.  New board is a different brand I had never used before + Intel Duo Core.

ZERO PARITY ERRORS!!!

 

Okay, problem fixed (thought I wouldn't say solved).  Last night changed motherboard & processor.  Ran parity check, which of course found thousands of errors.  Ran the subsequent check, and voila, no errors.  Same memory, power supply, cables, case, SATA PCI card.  New board had gigabit LAN, so ditched the PCI one I previously used.

 

What caused this problem is your guess.  Previous board was ECS (have to look up model when I get home) + Intel Duo Core processor.  New board is a different brand I had never used before + Intel Duo Core.

I would put the old motherboard into a PC loaded with some version of MS-windows.  It probably will not work any better, but the occasional errors will be masked by the frequent errors and crashes of the MS OS.  

 

Then again, since MS probably never accesses all the disks in parallel like unRAID, it might just work perfectly fine...

or...  you might just use that motherboard as a wheel chock for your car, and then determine how difficult it is to drive over it.... again and again... 

 

Congratulations on finally getting through a parity check with all your drives.  !!!!!!!!

 

Joe L.

Then again, since MS probably never accesses all the disks in parallel like unRAID, it might just work perfectly fine...

or...  you might just use that motherboard as a wheel chock for your car, and then determine how difficult it is to drive over it.... again and again...  

 

Joe L.

 

After years of building workstations for personal use & my office, I'm finally to the point that it's just not worth messing with hardware that is suspicious.  Due to the low cost @ Fry's Electronics for a board/processor combo, I'm gonna probably just trash the stuff.

 

thanks all for everyones help, and hopefully this helps others in the future.  I should also note it's nice to be able to swap a board & processor and not have to go through 1-2 hours of installing a Windows OS.  Maybe Microsoft should put their stuff on a flash drive 

After years of building workstations for personal use & my office, I'm finally to the point that it's just not worth messing with hardware that is suspicious.  Due to the low cost @ Fry's Electronics for a board/processor combo, I'm gonna probably just trash the stuff.

For the benefit of those who follow, you might want to post the exact motherboard brand and model.  That way, others will know to avoid it if possible.  (And entry in the hardware UN-compatibility list in the wiki is probably in order too)

Maybe Microsoft should put their stuff on a flash drive

 

  Next year the available USB drives might be big enough 

 

Don't toss the processor, it probably is fine.... if you insist, toss it my way 

 

 

WORKING:

Motherboard: MSI P6NGM-FI (onboard gigabit LAN & nVidia GeForce7)

Processor: Intel Core2Duo E7200

 

 

NOT WORKING (unknown of which component is the problem):

Motheboard: ECS P4M800PRO-M V2.0

Processor: Intel Core2Duo E4300

Trendnet Gigabit PCI LAN

 

All other variables consistent between the 2 systems.