I had a strange and scary thing happen today. I was trying to copy some data from my unRAID server over the network but it just kept timing out. Eventually I looked at the unRAID web interface and everything was green and reported as OK with no errors. I reboot the unRAID server anyway (it had been since the day beta13 was released) to see if it was some kind of a networking problem.

 

After the reboot I checked the unRAID menu and everything was still green, then I looked at the syslog via unMenu and it was full of red lines (mini heart attack), but still the array was all green and reported as working. From the logs I managed to determine which drive was causing the errors, I took the array off-line (this took ages) and removed the problem drive, I tested the drive in another machine and determined that it has indeed failed.

 

I had a pre-cleared drive (waiting patiently in a draw for 2 years) so installed it, re-started the array and it's now rebuilding at 115 MB/sec with 204 minutes to go (which I assume is normal).

 

The question that I have is why didn't unRAID detect the failed drive and report it to the web interface?

 

Log file attached.

syslog-2012-04-16.txt.zip

UnRAID only disables a drive on a write error. The drive was having read errors which unRAID should correct. Apparently the corrections were causing the transfers to time out.

That makes perfect sense, thanks.