Restic backup on local machine hangs while scanning files #3156
Comments
|
The stacktrace looks like several syscalls got stuck, at least two read and one open syscall are active for 68 minutes. Combined with the fact that |
MichaelEischer
added
the
state: need feedback
|
That's useful to know -- I could plausibly switch that disk to use a different filesystem without too much trouble. If I do, I'll update with whether the problem recurs or not. |
|
I switched to ext4 on the source disk and have made several backups without issue. |
|
Thanks for the feedback. So this looks like an issue of the underlying filesystem. For such problems it might be worth a try to set the environment variable |
Output of
restic versionrestic 0.11.0 compiled with go1.15.3 on linux/amd64
How did you run restic exactly?
restic -r /mnt/backups --verbose --verbose backup /mnt/ext
The first time I did this, restic ran for about 20 minutes, scanned a few dozen gigabytes, and then hung indefinitely, displaying the ETA increasing and the same two MP3s it was scanning as the last output, while making no progress. I attempted to kill it, but it didn't respond to even a kill -9. The only way I could get it to die is by
umount -f /mnt/exton the disk it was backing up (the first time failed with "target is busy"; the second time worked.) I then tried it again and it scanned for about 40 minutes before hanging with the same symptoms. htop showed that lots of memory was available and restic was no longer using any CPU. Plenty of disk is available.The second time, I sent a SIGABRT and retrieved a stack trace, which you can see at the bottom of this issue.
What backend/server/service did you use to store the repository?
My machine is running Arch Linux with the latest version of restic from the AUR. The path it was backing up, /mnt/ext, is a exFAT filesystem mounted from a local USB disk. The repository target, /mnt/backups, is an ext4 filesystem mounted from another local USB disk. Its cache is on a third local disk.
Expected behavior
Restic creates a new snapshot in the /mnt/backups repository.
Actual behavior
Restic hangs apparently indefinitely.
Steps to reproduce the behavior
So far this didn't happen when I was creating the repository (last week) but it's happened every time I tried to add a snapshot.
Do you have any idea what may have caused this?
Nope. Smells like some kind of deadlock.
Do you have an idea how to solve the issue?
Nope.
Did restic help you today? Did it make you happy in any way?
If I can get restic working I will be very happy to have an incremental backup tool I can rely on!
Output after SIGABRT:
The text was updated successfully, but these errors were encountered: