HELP!! My UnRaid keeps crashing all of a sudden...

[jimwhite]

After many, many months of trouble free service, my server is crashing....  It's built around the original Intel motherboard... does this error screen give any clues?

 

[limetech]

It looks like it's running out of memory.

 

Two questions:

1. How much RAM is installed, and

2. If you are using User Shares, do you have some very large directories which span multiple disks?

[jimwhite]

It has 1 gb of DDR333

 

Yes, there are some VERY large directoriy structures shared that way, BUT, nothing on those trees has changed for months.... it just started crashing a few days ago.... the day after a long power-failure in which my UPS battery ran down.

 

The message on the screen yeaterday had to do with interrupts...

 

maybe I have a RAM problem?

 

[limetech]

Might be a defective RAM, but not likely.  Please send me your system log:

 

From telnet type:

 

  cp /var/log/syslog  /boot/syslog.txt

 

Now the file syslog.txt should show up in the Flash share where you can copy it.

[jimwhite]

Okay.... will do that tonight when I get home....  thanks! 

 

 

Might be a defective RAM, but not likely.  Please send me your system log:

 

From telnet type:

 

  cp /var/log/syslog  /boot/syslog.txt

 

Now the file syslog.txt should show up in the Flash share where you can copy it.

[jimwhite]

okay... zipped syslog attached.... 

 

[jimwhite]

Okay.... here is the most common error screen....

 

 

[limetech]

"Unable to handle kernel paging request..." is the culprit here.  It's running out of memory.  Why exactly this is happing now is a mystery (but with enough work could probably figure it out).  Please upgrade to 4.0.

[jimwhite]

upgrading at this stage sounds a little scary.... it's been working so well for so long.... I remember that when I first put this thing together, I found the Intel motherboard very "picky" about memory.... it wouldn't work very stably at all with some very good Corsair memory, even when underclocked.... I put in the current set of Samsung DDR333's and after fiddling with the memory settings it was working fine.... but maybe something got reset during the powerloss.... at any rate, I pulled one of the 2 512mb dims and it's now past 50% of the parity check... before it would never get past 2%.... I think this memory/motherboard combo doesn't like dual channel?  oh well.... some experimentation is in order

 

as a side note, I see that the parity check has turned up 4 sync errors... maybe not surprising considering all the crashes over the past few days... is there anything that will tell me what these sync errors were?

 

[limetech]

Sync errors are cases where the stored parity data didn't match newly calculated parity for the stripe being checked.  System crashes do explain these errors.

 

Regarding picky memory - yes this is a relatively common problem.  Can also be caused by marginal power supply.

 

Regarding upgrading - no guts, no glory    Seriously, you can upgrade to a new Flash relatively easily and go back if things don't go right.  If you need a license key for this, send me an emal: [email protected].