"Error occured in launcher loop"

[Jabsu]

After one of the latest updates, I've started to get these:

[2017-04-20 18:59:58][ ERROR][overseer] Error occured in launcher loop. Traceback (most recent call last): File "...\overseer.py", line 364, in launch await self._launch(update_spawns) File "...\overseer.py", line 396, in _launch for point, (spawn_id, spawn_seconds) in spawns_iter: TypeError: 'NoneType' object is not iterable

The scanner will always forcibly close after these. Maybe something to do with the database? I'm using MySQL and made a new database just to be sure, but the problem still exists.

[leald95]

Same thing is happening with me while using Postgres, thought I was the only one.

[Jabsu]

I removed the spawns.pickle file and now the problem seems to be gone.

[leald95]

I changed a few config settings around that had to do with spawns and altitude and what not, and it seems to be working for now, but I will try that next if it happens again. Thank you.

[patienttrainer]

Same here. Same setup. Same solution.

[steffenkloster]

Seems to happen again and again. Deleting spawns.pickle only solves it temporarily

[Jabsu]

I haven't had the problem anymore. I'm not familiar enough with the code to say for sure, but I guess it has something to do with unknown spawns that are out of reach (outside of map boundaries, e.g. when using polygons). By removing the spawns.pickle file, those out-of-reach spawns are gone from the workers' mind, so they can concentrate on the spawns that matter. This will help (or at least in my case it helped) with skipping, timeouts and clustering too.

Anyway, I hope this will get some attention, as it's pretty critical.

[ghost]

Has anyone tried running scan.py --bootstrap instead of deleting spawns.pickle? I had a 'NoneType' object is not iterable crash yesterday, and that solved it. I don't remember whether it was the same place in the code as described in this bug report, though.

[ghost]

This crash has been happening to me a lot since the increased event spawns arrived. Running scan.py --bootstrap does temporarily work around it without having to delete spawns.pickle.

[ghost]

I believe the cause is this line in MoreSpawns.add_known():

# add so that have_point() will be up to date
self.known[point] = None

The overseer code expects the known dict to contain (spawn-id, spawn-time) tuples. When it finds None instead, it breaks.

[DavePlater]

I was getting this frequently up until a few weeks ago when I updated everything (requirements and monocole). The fix had always been to delete the pickle.