web2py scheduler failing on PythonAnywhere since Dec 1

Question

Since Dec 1, I can no longer schedule tasks in the scheduler for my web2py (Version 2.14.6-stable+timestamp.2016.05.10.00.21.47) app on PythonAnywhere.com; the task log is a LONG permutation of the following 3 lines:

ERROR:web2py.scheduler.giles-livetask1#7087:    error popping tasks
ERROR:web2py.scheduler.giles-livetask1#7087:Error retrieving status
ERROR:web2py.scheduler.giles-livetask1#7087:Error coordinating TICKER

interspersed every hour with something like

Failed to acquire lock, task must already be running
2017-12-06 08:04:12 -- Completed task, took 6.00 seconds, return code was 0.

(Supposedly they monitor SO, but since I have had no response via direct query nor via their forums, I'm not holding my breath.)

I have tried restarting the scheduler, to no (helpful) effect. So my question is: how do I get this scheduler working again?

Just a note from PythonAnywhere here -- we answer all questions on the forums and over email within 24 hours, normally much faster during the business day. As the original forum post the above refers to was at 9pm our time, it didn't get answered until the next day. We'll be helping debug the problem on our forums, so here's a link to the thread: https://www.pythonanywhere.com/forums/topic/12172/ — Giles Thomas, Dec 07 '17 at 17:42
Follow up: I *never* received feedback from the feedback I posted, which I first did the morning of my forum posting. And while I did get feedback on the forum, it only addressed restarting the scheduler which, as the question above makes clear, did not address the problem. — Scott Hunter, Dec 11 '17 at 14:40
Scott -- I'm really confused now. There has been a question pending a response from you on the forums since December 7 -- check the forums link above -- and we sent you an email in response to your feedback message the same day. If you didn't get an alert about the forum post, and didn't receive the reply to your feedback, could you check your registered email address on PythonAnywhere? Perhaps our messages are going to an email address you're no longer checking. — Giles Thomas, Dec 11 '17 at 20:14

score 0 · Answer 1 · answered Dec 11 '17 at 14:43

Poking through the scheduler source and deeper diving through the logs suggested that the SQLite database was locked; https://stackoverflow.com/a/15271260/535275 provided the rather simple steps (basically, make a copy of the DB file then rename it to replace the original) to rectify things.

web2py scheduler failing on PythonAnywhere since Dec 1

1 Answers1