1168957 - Check that crontabber doesn't start the same job twice

Status: RESOLVED FIXED

(Socorro :: General, task)

Description

[Peter Bengtsson [:peterbe]]

Ideally we should be able to do this:

./crontabber.py --admin.conf=myjobs.ini &
./crontabber.py --admin.conf=myjobs.ini 

I.e. two "parallel" crontabber instances.

We used to determine which jobs to do on based on next_time but we now also have the ongoing state which should prevent instance number >1 to not run.

FTR, we're not worried about starting two instance on the same server but possibly (accidentally) starting up more than 1 EC2 instances using the same config.

Comment 1

[Peter Bengtsson [:peterbe]]

Turns out, it's possible to start a second instance of crontabber whilst a first instance is in the midst of running a job :(

https://github.com/mozilla/crontabber/issues/68

Comment 2

[Peter Bengtsson [:peterbe]]

The feature is now built into crontabber. It's now resilient to running concurrently against a central postgres store without trampling over itself by doing the same job twice in parallel.