SABnzbd Forums

I've got Mylar (v3) set up to connect to SAB (v3).
Previous versions worked perfectly fine, I'd add comics or set them to download, SAB would get it, download it, and using the ComicRN script, would pass it back to Mylar for renaming/metadata tagging and moving to its correct location.

However the recommendation is to use Completed Download Handling, similar to Sonarr, etc.

Now when I kick off a comic download, SAB still gets it from Mylar and downloads it. But it doesn't get back to Mylar. It just sits in the completed folder for Mylar (as set by the categories)
And looking at the log, it seems that Mylar is searching for that downloads NZO_ID, but can't find it anymore, so it can't pull the file back into Mylar (the Mylar dev confirmed that this is what seems to be happening)

So what in SAB's config would be causing the downloads it processes to change its NZO_ID?
I've got History Retention set to 'Keep All', and 'Check Before Download' and 'Detect Duplicate Downloads' are also off (at Mylar devs recommendation).
I know I'm missing something, but I've tried so many different permutations of options, switches, etc, and I'm real lost!

Interestingly enough, in looking at the SAB logs, I can see that:
[nzbstuff:1759] [sabnzbd\nzbqueue.pyo.cleanup_nzo] Purging data for job

And its after this point that Mylar can no longer see the finished download.

Any idea what might be causing this mid-process clear?

Could you share more of the log? Maybe it's a bug in SABnzbd 3.

Abolsutely!
Here are two clean logs that I collected and added to the Mylar git issue on this:
You'll have to remove the square brackets to make it work

github[.com]/mylar3/mylar3/files/4626297/mylar.log
github[.com]/mylar3/mylar3/files/4626299/sabnzbd.log

Hmmm, strange.
I see until the end "SABnzbd_nzo_4lk9dbta", just as it was added.
And then mylar searches for the whole history: Are you sure the NZO-ID changed? If you do a manual call to the history API.