Hi All,

 

Well I am getting more concerned with these growing problems. I has ran rebuild tree twice on disk2 and ran smart on all the drives. Here is a syslog from last nights errors. Help Please.

 

Thanks,

 

  Dave

 

http://www.pastebin.ca/1269640

 

 

 

Since the parity check starts before the drives are even mounted, there must have been something really wrong on the previous run.  What can you tell us about the state of the server when it was last shutdown or rebooted?  Do you possibly have a syslog from that session?  You may want to install the powerdown script, and use it to power off, because it automatically saves a final syslog.

 

Something sounds familiar in your comments, do you have another thread about this?  If so, it would be better to continue in that one, otherwise helpers have to ask for history and more detail and your equipment all over again.

 

Some time after the parity check finishes, at 7:20PM, an action was initiated, related to spin down, and 5 seconds later, a series of exception Emask error sequences (type: frozen, timeout) began for 9 of the 11 drives.  I believe the action was the use of the Spin Up button, as the syslog messages, the unRAID version, the error sequences, and type of errors, and the resets issued, are all consistent with a Spin Up failure.  I don't know if this problem is related to the other issues you have been having, but it may be a cause.  Some where, and I don't remember where, I have issued a warning against the use of the Spin Up button when running unRAID v4.3.3.  With v4.3.3 and possibly a few versions earlier, Tom was trying to improve the reliability of the Spin Up command, and a number of users, like myself, were testing various alternatives.  We thought we found a better way, and Tom added it in v4.3.3, and only later I discovered that although it worked great, it often resulted in the errors you have, and also caused a number of soft and hard resets, that can sometimes temporarily (until next reboot) cause loss of access to one or more drives.  I don't know how many systems are affected, but certainly not all, perhaps only a few, mine being one too.  Your system looks particularly vulnerable.  I would recommend that you NEVER click the Spin Up button, until you have upgraded or downgraded from v4.3.3.  In this particular case, it looks like all of the drives were recovered, but that may not always be the case.

 

What I *think* is happening, is that this method of simultaneous issuance of the spin up command to each of the drives, causes the system to 'lock' on the first drives, as if it is waiting for each drive to complete spinning up, before it can continue.  But the commands have been issued to all of the drives, with a timeout period of 5 seconds, which is not long enough, and they then timeout and begin trying to reset the drives, thinking something is wrong.  The only 2 drives of yours that did not have these errors were the first drives on 2 independent channels.

 

Note to others trying to figure out how to download from PasteBin:  look for the word Raw in the top right, and right-click on it and use the Save link as ... option, or your browsers' equivalent.  You will have to rename it, but that will give you the original file intact, without all of the header garbage, html codes, AND the aggravating line wrapping.  I don't know why they don't have a Download button or link.

Rob,

 

Thanks for the detailed reply, I think this all started when I installed a Super Micro 5-3 bay chassis. with Parity - Disk 4. I also replaced disk 1 and 2 with Seagate 1TB drives in sequence. later found to be a DOA drive. I replaced with a Samsun 1TB and rebuild data again, following this I started to get Parity errors and joe suggested that disk 2 had a corrupt files system so I ran reiserfsck on disk2 and in two instances running not a week apart I needed to rebuild the tree and it found many lost and found files etc. That is from this post.

 

I do not have a syslog from the last session. I think what happen is I reconfigured my network with a new workgroup name and router and in the process forgot to log into the server and shut it down via the console. So I made all the changes and then had to issue a shutdown -h now at the server upon repowering the power died for a split and caused another reboot of the server, but I think this happened as the system was in the process of booting and mounting the drives.

 

http://lime-technology.com/forum/index.php?topic=2662.0

 

I am not sure about the spin up issue, I never use that function at all, but I noticed with this latest reboot that the drives seem to be spinning down.

 

Can you post a link to the power down script please.

 

 

 

Dave

Hey All,

 

Juat an update, I have ran reiserfsck on all my drives including the problem file structure drive disk2 and they all report no corruption. I have also ran a short test on all 11 drives and have no reallocated sectors on any of the drives and they all pass the short test fine.

 

I am going to capture a syslog now and then will run another parity check overnight.  I have two new 1TB drives to install but not until I get my errors back to zero on the parity drive.

 

dave

Well all is well. it just finished a Parity Check started this morning before I left for the shop and good news - A bif fat GOOSEEGG for errors... ;D

 

Dave