PDA

View Full Version : Server crashes when accessing SUS directory




jbaileypro
Jan 11, 2013, 08:05 AM
We have had our server running fine for a good 18 months. One day however the server locked up to the state that I was able to move the mouse and move windows however menus etc wouldn't load. I contacted apple care support and sent off my logs but haven't heard back from them.

This ended up happening multiple times and I was able to pinpoint that it was when our backup software (carbon copy cloned) that ran at the time of each crash. Each time it was when it was backing up a file inside the sus directory (software update server directory). I also ran a HD program to check the disk for bad sectors and this again crashed in the same directory. Now I have removed the directory from the backup and the server hasn't crashed. I have stopped the software update service too (before the backups ran).

I am worried to why this is happening. All the drives report as fine in disk utility. The backup are running but I don't want it to lock up again for a more important directory. The only way to rectify the crash is to power cycle the server and all services are unresponsive when the crash happened.

I also tried starting the mini in target disk mode and using a separate mac to clone the drive. The disk disconnected when I reached the same sus directory. I had to use disk utility to manually remount the drive as it showed as being disconnected (with the error message to go with it).

Any help or ideas would be much appreciated.

Jonathan



jared_kipe
Jan 12, 2013, 02:17 PM
I would suspect that your HDD is going out.

Two things, download black magic HDD speed test utility and get an idea of the speed at which your HDD is reading and writing.
I recently had a RAID fail. The first indication was that a Windows VM refused to boot because of an IO error, but I passed it off as some windows corruption on the vm's HDD file. Secondly, a scheduled rsync backup was taking a LONGgg time to complete. In my case I had non-raided partitions on each disk, so a couple of minutes in Blackmagic showed me which disk was able to read at 150MBps and which one was only doing 10MBps. Sometimes drives limp to their deaths.

If your speed tests are inconclusive, try copying that directory with ditto. Ditto has some pretty nice error handling built in and 'usually' is quite graceful under IO errors when compared to other programs. Open up terminal and type in 'sudo ditto -rsrc -V /path/to/sus/ /tmp/sus/'

If the files wizzing by indicate an error, I'd probably replace the drive. (You could try to delete that directory, but I don't know that I would bother)