Could someone have a look at my syslog for me and advise me if anything / what is wrong.

I started to preclear 4 samsung HDD's yesterday evening and when I got up this morning I could not access the Tower over the network.

I rebooted my main pc and the router and still could not access tower. I had four tty sessions running and it looked to me that the pre clear sessions had completed.

So I exited three of the four sessions and typed reboot and the server rebooted. I now have a red ball against drive 11.

I have powered down and rebooted again and drive 11 is still redballed.

I have attached syslog . I cannot see anything obviously wrong in the syslog when i look.

Disc 11 is sdg

 

Thank you

 

I tried to attach a few screen prints from pc but files were to big

syslog-2010-12-05.txt

A "write" to the drive failed.

 

See here:

http://lime-technology.com/wiki/index.php?title=FAQ#What_does_the_Red_Ball_mean.3F

Thanks Joe L.

 

Here is a smart report on that drive..it looks okay? Can I use The 'Trust My Array' Procedure I have not written anything to the array or changed anything and I have not added in the four drives yet that I have pre cleared.

smart.txt

Thanks Joe L.

 

Here is a smart report on that drive..it looks okay? Can I use The 'Trust My Array' Procedure I have not written anything to the array or changed anything and I have not added in the four drives yet that I have pre cleared.

For the drive to have a "red" ball, a "write" to it failed.  Even though you did not remember writing to the disk, it was written to.

It might be something very simple, like the mount status, or it could be part of the file system structure, or it could be part of a file.    You can only guess.  From what you said your server crashed and had to be rebooted.  All we know for certain is that parity will probably have a change that did not get written to the disk.  That change will be lost if you use the "trust" procedure as it would trust the disk data not the parity.

 

I Strongly suggest you stop the array power down re-seat the power AND data connections to the drive before anything else.  Be careful to NOT dislodge the connections to any other drive, otherwise you'll have two drive failures and not be able to start the array.

 

As long as you are aware that if you use the "trust" procedure and risk allowing possible damage to the file system,  you make an informed decision.   Whatever you do, you'll probably want to verify all is good by performing a file-system-check on the drive after you are back up running, before writing anything more to that same drive.

 

Personally I'd not use the "trust" procedure.

 

Joe L.

Whenever something unexplained happens in your unRaid server, your first priority should be to save the syslog. It is your only chance of figuring out what happened and why. If you reboot it is gone and you wind up just guessing what could have happened.

 

If a drive is kicked from the array, yet on reboot it looks fine that usually means you have a bad or loose data or power cable to that disk.

 

At a minimum I would resecure.these connections before running the Trust procedure.

Okay...here is what i have decided to do. I am copying the data from disk 11 to disk 10. I know the data is not actually being read from disk 11 rather it is from parity. When the copy is complete I will remove disk 11 and run some checks on it on my windows pc. I will then pre clear the disk and see what the pre clear results come up with.

I do not have a 2tb HDD to replace disk 12 at the moment.

The copy process is 50% complete.

 

I use each hdd as an individual drive, I do not share data across all hdd's so this exercise should be relativity easy to run with.

 

I will report back when I have run tests.

The pre-clear is not needed...  You can restore full parity protection to your array by

 

Stopping the array

un-assigning the failed disk.

Powering down.

checking the power and data connections.

Powering up.

Starting the array with the disk un-assigned.

Stopping the array.

Re-assigning the disk to its slot in the array.

Starting the array.   It will re-construct the "correct" contents, as read from parity and the other disks to it.

This re-construction will take about the same time as the initial parity calculation.  You should be back parity protected in less than 12 hours.

 

The procedure you described will probably take much more time and leave you without parity protection for that longer interval.

 

Joe L.

The pre-clear is not needed...  You can restore full parity protection to your array by

 

Stopping the array

un-assigning the failed disk.

Powering down.

checking the power and data connections.

Powering up.

Starting the array with the disk un-assigned.

Stopping the array.

Re-assigning the disk to its slot in the array.

Starting the array.   It will re-construct the "correct" contents, as read from parity and the other disks to it.

This re-construction will take about the same time as the initial parity calculation.   You should be back parity protected in less than 12 hours.

 

The procedure you described will probably take much more time and leave you without parity protection for that longer interval.

 

Joe L.

 

Doing that now...thanks guys

Everything ran okay Disk 11 is reconstructed and back on line. Parity completed with no errors. I'm wondering if the problem was occurred whilst I was pre clearing four hard drives as when i added the four drives after the rebuild of drive 11 finished two had to be cleared and a full format was performed on all four disks. So it looks like the pre clears did not complete. Anyway here is my syslog if anyone fancies a look.

My two Sharkoon fan holders arrived on Saturday from Italy.  http://lime-technology.com/forum/index.php?topic=8784.0 So i can finish building the server now. Everything is just spread out on the desktop at the moment. I will put some pics up in the lounge when the build is complete.

 

Syslog..  Can anyone recommend a file hosting site please.  My syslog is to big to post here and the one i just tried has loads of crap popping open windows etc!!

Personally I'd not use the "trust" procedure.

 

Hmmm ... I ALWAYS do the trust procedure for something like this.  Must have done it 4-5 times over the years.  The trust procedure leaves parity intact and the drive in question is not recreated from scratch.  Of course if I had written to the disk, especially something pretty big, I'd do the rebuild but onto a difference physical disk.

 

I have never rebuilt a disk back onto itself.  I'm always afraid the drive rebuild is going to fail and I'd be stuck.  Much better to have the disk, even if it is minus a couple of updates. 

Syslog..  Can anyone recommend a file hosting site please.  My syslog is to big to post here and the one i just tried has loads of crap popping open windows etc!!

 

Many use Pastebin

Thank you mbryanr for the link.

I have now installed 16 hdd's into my server and they are all up and running but I would like to run a few checks on HDD 11 that had the write errors and caused unRAID to disable the disk.

Can any of you guys suggest a tool / command that will enable me to run some tests on HDD 11. I plan to move all the files on that disk to another disk first.

I do not use shares of any kind, I select manually which disk I want to put folders and files on.

 

Thank you