Getting error when backing up, but cannot find log info

I’m using Duplicati to backup my shares on my unRAID server. I’ve set up Duplicati to run every night. But it hasn’t run for a while now, and the only error message I get is a popup saying “Some kind of disk I/O error occurred disk I/O error”. The latest log entry I’ve got is from when Duplicati last ran successfully (Feb 14th). And I get no info in the Docker container console log when trying to run the backup manually either.

Any idea what I can do to troubleshoot this?

I noticed now I actually get a BTRFS error in the unRAID system log when trying to start the backup process in Duplicati.

I’ve seen these error messages before, but thought it could be a parity issue. I’ve also had issues with attached USB devices on my server, so I’ll admit that these error messages appearing not so often haven’t caught my attention until now.

I don’t know enough about Duplicati yet, but its either Duplicati or your filesystem/drive. I know some stuff isn’t compatible with btrfs but don’t know if its possible here.

Looks like btrfs corruption. Duplicati itself is filesystem agnostic - it just asks the OS to read files. I use btrfs on my NAS and run Duplicati without issue.

You might need to do some sort of repair on the filesystem. While my NAS has btrfs I’ve never needed to do a manual repair. (I have a scrub operation scheduled for once a month, but do nothing beyond that.)

FWIW there’s information on the Web about messages like yours, but understanding them is beyond me.

I just noticed that if Duplicati shows a notification, for certain info it looks to be required to go to about → show log → live → select the level (in this case it would be error).

I was getting a huge amount of notifications with no info at all and that was the only thing that showed me what was happening.

A completely empty popup box? Can you post a copy of the box?

What often happens on backup errors especially, is the failure goes into the server log instead of the log files of a backup job, which probably explains the OP (please look), but sometimes details are in live log.

Those are reproducible if eg email alert isn’t setup right.

It was a lot of notifications that just said a warning has occurred (or something like that) and the links did nothing, dismiss did nothing. Didn’t know what it was but something had to be somewhere :slight_smile:

That needs a fix. The links should go to the log or something.

This is very vague. It probably merits its own forum topic or better yet file an Issue with full information.
The forum has no facilities to track things that need a fix, whereas putting them in Issues gets in line…
There are certainly some known issues with logs (I assume their fix is hard, or they’d already be fixed).

Probably, but I’m not concerned with it right now.

This sounds related

Could be, but “needs a fix” post is vague. Maybe putting known log issues out will jog memories.

is odd. The combined logs issue would do something, if you ask a popup box to show you the log.
Problem is it takes you to the job log even if the error was in the server log, but that’s not “nothing”.

might be this one where it does something, but the dismissal is off the visible scrolled screen area.

Interface blocked by hundreds of old warnings, “Dismiss all” does not work

Do you happen to know if the invisible dismiss or the mass-dismiss-only-gets-some has an issue? Certainly things can’t all be fixed at once (or maybe ever), but at least these things can get in line…

There are more. Finding log info can involve looking many places, sometimes, including the live log (because it’s not in the job log or the server log…). I’d love it if logging could be improved somehow.

You’re right. My problem was BTRFS corruption on my cache SSD drive (where I store the docker appdata). I ended up backing up the cache drive, formatting it to XFS instead, and copying the backup data back again. Had to recreate the Duplicati database, but now it’s up and running again.

Thank you for the help!

1 Like