I have an Ubuntu 14.04 LTS I5-based system that hosts three VM's and has been running happily for about a year. I recently set up a USB drive to back up the VM drives and wrote a simple script to shut down the VM's, copy the drives to the backup, and restart the VM's every Monday morning at 3AM. Simple eh?
Well, not quite. Approximately half the time, the job fails with one of the cp operations in the script terminally hung in "D" state (ps -ax / uninterruptable sleep) that can be traced back to a kernel panic (shown via dmesg) wherein the cp task "was blocked for 120 seconds."
I had previously researched this error and found configuration settings that had claimed to resolve this issue as noted here that directed the addition of these two lines to /etc/sysctl.conf:
vm.dirty_background_ratio = 5
vm.dirty_ratio = 10
I did this, rebooted, and backups ran succesfully two weeks in a row. I thought the issue was resolved, but it recurred again last night.
The files being backed up are 50+ GB files, going to a 2TB backup drive. I've had zero other issues with this box since I built it one year ago - only since this backup drive was introduced has this issue manifested itself. So I'm inferring the backup drive must at least be considered as part of the problem.
The only way to solve this issue is to hard-restart the system, which is never a desirable option. Has anyone any insights into this issue? Is it a linux kernel issue, or could I be dealing with a temperamental USB backup drive?
Any other suggestions welcome and appreciated.
