Frequent Crashes - unknown cause

[mcai3db3]

After removing the dirty cache drive, it hasn't crashed yet. Been about 4 days. Still evaluating.

 

Gonna add a cache drive (a different drive) soon, see if issues arise then.

 

Not touched voltages yet. Scared.

[RobJ]

Going by the SMART report, the Cache drive looks fine.  Are there other reasons why you are suspicious of it?  There was a corrupted sector around 440 operational hours ago, but that was recovered.

 

The thing that is not OK is that "irq 16: nobody cared" message, and especially when it is followed by "Disabling IRQ #16".  Can you attach that particular syslog?  If it matches the syslog you did post, then that would have meant the loss of access to Disk 4 and Disk 5 (as well as of some USB ports, probably not in use).  A spurious interrupt like that is serious, could possibly mean a bad motherboard, or bad controller, or possibly a buggy application, or ... don't really know.  Motherboards based on nForce chipsets are notorious for spurious IRQ7's, but if they cause problems, it is easy to reserve that interrupt in the BIOS, and thereby disallow its use (sidelining it).  I really don't know about yours yet, or whether it is even related to your reboots.  I don't think IRQ 16 can be 'reserved'.

 

Hopefully we will know more from another syslog.

[mcai3db3]

I've not got that syslog anymore, but I can tell you exactly what I can do to cause such an error...!

 

Earlier today I will have got that message, it was rebuilding a drive I removed, onto a new drive. So the array was 'Starting' at this point. Then I went into the 'Shares' tab of the unraid management page. I made a change to one of the shares, adding a disk to a share or something like that. At this point when that page refreshes, all but one of my Shares shows on the Shares setup page, the rest disappear. Once the array is started up, only the one remaining Share is findable on the network. So, at that point I decide I want to reboot the thing, so I can have my Shares back!... OK, so I disconnect everything from the server, and then I try and stop the array.

 

Not so fast.

 

UnRAID says no. It basically just gets stuck on Unmounting for multiple drives, until I end up powering down with UnMenu.

 

When I restart, all is fine.

 

I get that IRQ message during that whole affair. Right now I'm in no mood to be fiddling around with the box, it's been working for 4/5 days and I just don't want to mess with it unless it proves itself to be an issue.

 

 

 

Oh, and the only reason I was suspicious over my cache drive was due to frequent inability to unmount said drive. But that could've been caused by anything. It was also a convenient way to remove my 500GB Cache and replace it with a 2TB data drive that I just bought  (and downgrade a 1TB data drive to cache duty).