I am running version 4.4.2 with 13 data drives (mix of sata and pata).  Recently after running for about a day or two, the tower becomes unresponsive (cannot telnet, no web server).  Once unresponsive, I have to hard reboot, which causes a parity-check, which shows no errors (after about 14 hours!).  Then a day or two later the cycle repeats.  Unfortunately, once the issue occurs I can not get access to syslog to look for any clues.

 

Any ideas?

I am running version 4.4.2 with 13 data drives (mix of sata and pata).  Recently after running for about a day or two, the tower becomes unresponsive (cannot telnet, no web server).  Once unresponsive, I have to hard reboot, which causes a parity-check, which shows no errors (after about 14 hours!).  Then a day or two later the cycle repeats.  Unfortunately, once the issue occurs I can not get access to syslog to look for any clues.

 

Any ideas?

 

In the unRAID Add-on section of the unRAID wiki you will find a section for the Power down package.  Get that installed and then do a control-alt-delete next time the server freezes.  It should save a syslog to the server before it does its shutdown/restart.

and ask yourself this simple question :

Why is my server freezing ?

Is it unRAID ? Is it my PSU ? Mobo ? Bad cable ? Add on card ? HD ? etc ... and change it.

Then a day or two later the cycle repeats.  Unfortunately, once the issue occurs I can not get access to syslog to look for any clues.

 

login as root through the console.

then type tail -f /var/log/syslog pess enter and leave it there.

 

if the machine freezes go back and see if this screen updated with messages. It may provide a clue.

Press enter, if it scrolls up, then the hardware itself probably has not frozen.

 

Press CTRL-ALT-F2 or F3, get another login session and see if you can login.

 

If it's just the network, the machine will be responding, but no network transfer can occur.

 

My powerdown package does capture syslog upon reboot,  so that may provide assistance to others for review.

 

 

Are you using sleep / S3 or is the server up all the time?

 

Random question, I know - but the reason I ask is that my machine was exhibiting the same behavior.  Normally associated with moving a lot of data to / from the server. 

I'm just about to begin to troubleshoot, and the first place I'm going to look is the sleep / s3 function because the behavior feels like it started around the time I started fiddling with S3.

 

My original post / description is here: http://lime-technology.com/forum/index.php?topic=4484.msg41320#msg41320

Are you using sleep / S3 or is the server up all the time?

 

Random question, I know - but the reason I ask is that my machine was exhibiting the same behavior.  Normally associated with moving a lot of data to / from the server. 

I'm just about to begin to troubleshoot, and the first place I'm going to look is the sleep / s3 function because the behavior feels like it started around the time I started fiddling with S3.

 

My original post / description is here: http://lime-technology.com/forum/index.php?topic=4484.msg41320#msg41320

 

I did take a look at both syslogs back when you posted them, but had no time, and did not see anything particularly wrong, except that there were a LOT of scary messages and exceptions in connection with the sleep/wake events.  Taking a longer look now, the first syslog does exhibit the loss of your flash drive on the last wake up.  It appears to have awakened successfully the first time, in both syslogs, and several more times in the first syslog, but finally the USB port could not be recovered.

 

The problem with the flash drive in no way that I can see, explains running out of memory, which is the most likely explanation for the behavior you are seeing.  We really need the syslog, plus it may be useful to monitor memory usage, and whether processes are being terminated.

 

I have to say that the sequence of device recovery after awaking is very off-putting, and it does not surprise me in the least that there was a failure.  I'm glad that it seems to work most of the time, but does not give me any confidence, at all.  The recovery does not look like a normal bootup, but more like a series of attempts to recover all of the *failed* devices, with many resets and exceptions.  From what BubbaQ has been saying, S3 should be improving with each new kernel release.