"_FAILED_xyz" folder left in main download directory - why?

[Cantello]

I noticed that since not too long ago (some months?) failed download folders with unpacked files are left in the main downloads directory. Any idea why they are left there and not in "incomplete downloads"? Any idea on how to automatically remove them? I have "remove failed downloads" running daily via the scheduler but this apparently removes the failed jobs from sab's queue but not the folders.

I am using v4.03 via docker on a debian machine. Permissions should be good as sab can normally write to all folders it has access to. There is nothing in the logs that pertain to these folders, neither that sab created them, nor that it tried to delete them and failed.

What can I do?
Thanks! :-)

[safihre]

It is only when we already started unpacking that this folder is left. Just in case users want to do something with the partly unpacked files.
Only if the download already fails during the downloading, it is only present in incomplete.

[sander]

There is nothing in the logs that pertain to these folders, neither that sab created them, nor that it tried to delete them and failed.

Really? Logging set to +DEBUG?

Because I see this in my logging:

Code: Select all

sander@haring1:~$ cat .sabnzbd/logs/sabnzbd.log* | grep FAILED

2022-08-04 08:49:47,301::DEBUG::[filesystem:866] Renaming "/home/sander/Downloads/complete/_UNPACK_..xyz." to "/home/sander/Downloads/complete/_FAILED_...xyz"

... which confirms what Safihre says: it was already in complete, and stays there.

[Cantello]

Really? Logging set to +DEBUG?

You are completely correct, this showed after setting log level to debug.

So, currently there is no way to automatically remove these folders, right? Maybe I will just implement a small cronjob with a batch script to remove these...

Thanks for the info!

[safihre]

You could do that, but it's actually strange that it can get this far. A broken job would have been caught by par2 verification.
Do you maybe have Post process only verified jobs turned Off in Config Switches?
And why are these jobs failing? What error is reported?

[Cantello]

You could do that, but it's actually strange that it can get this far. A broken job would have been caught by par2 verification.
Do you maybe have Post process only verified jobs turned Off in Config Switches?
And why are these jobs failing? What error is reported?

Post process only verified jobs is switched on as per default. The error is reported from unrar: "packed data checksum error". The relevant part of the log (this time with debug switched on) is:

Code: Select all

Extracting from /incomplete-downloads/xyz/<HASH>fe86832.part1.rar

Extracting  /downloads/_UNPACK_xyz/xyz.mkv

Extracting from /incomplete-downloads/xyz/<HASH>fe86832.part2.rar

...         xyz.mkv
xyz.mkv : packed data checksum error in volume /incomplete-downloads/xyz/<HASH>fe86832.part2.rar
2023-09-28 02:26:19,911::INFO::[newsunpack:310] Unrar finished on /incomplete-downloads/xyz
2023-09-28 02:26:19,934::DEBUG::[filesystem:501] build_filelists(): joinables: []
2023-09-28 02:26:19,934::DEBUG::[filesystem:502] build_filelists(): rars: []
2023-09-28 02:26:19,934::DEBUG::[filesystem:503] build_filelists(): 7zips: []
2023-09-28 02:26:19,934::DEBUG::[filesystem:504] build_filelists(): ts: []
2023-09-28 02:26:19,935::INFO::[postproc:437] Unpacked files []
2023-09-28 02:26:19,935::INFO::[filesystem:330] Checking if any resulting filenames need to be sanitized
2023-09-28 02:26:19,935::INFO::[postproc:441] Finished unpack_magic on xyz
2023-09-28 02:26:19,937::DEBUG::[filesystem:610] Applying permissions 0o775 (octal) to /downloads/_UNPACK_xyz
2023-09-28 02:26:19,938::DEBUG::[filesystem:610] Applying permissions 0o664 (octal) to /downloads/_UNPACK_xyz/xyz.mkv
2023-09-28 02:26:19,941::DEBUG::[filesystem:876] Renaming "/downloads/_UNPACK_xyz" to "/downloads/_FAILED_xyz"
2023-09-28 02:26:19,945::DEBUG::[filesystem:919] [sabnzbd.postproc.rename_and_collapse_folder] Removing dir /downloads/_UNPACK_xyz
2023-09-28 02:26:19,946::INFO::[nzbstuff:1833] [sabnzbd.postproc.process_job] Purging data for job xyz (delete_all_data=False)
2023-09-28 02:26:19,946::DEBUG::[articlecache:152] Purging 0 articles from the cache/disk
2023-09-28 02:26:19,961::INFO::[notifier:123] Sending notification: Download Failed - xyz (type=failed, job_cat=tv)

[safihre]

What about the part before, where it does the verification?
Maybe you can just send me the whole log at safihre@sabnzbd.org?

[Cantello]

Maybe you can just send me the whole log at safihre@sabnzbd.org?

Done, thanks for your offer!