1192380 - manually run truncate job on prod

Status: RESOLVED FIXED

(Socorro :: Database, task)

Description

[Robert Helmer [:rhelmer]]

Bug 1191006 truncates at 2 weeks, but this job only runs once per week and we're getting too low on disk to wait.

Comment 1

[Robert Helmer [:rhelmer]]

Testing this on staging - I don't think the crontabber job is actually working:

envconsul -prefix socorro/common -prefix socorro/crontabber /data/socorro/application/socorro/cron/crontabber_app.py --job=truncate-partitions --force
(...)
2015-08-07 19:42:23,942 DEBUG - MainThread -  - MainThread - successfully ran <class 'socorro.cron.jobs.truncate_partitions.TruncatePartitionsCronApp'> on 2015-08-07 19:42:23.409484+00:00

All this job should do is run "select truncate_partitions(2)" but it returns way too quickly, I don't think it's actually running this.

If I psql into the stage DB and run "select truncate_partitions(2)" then it seems to work correctly. I think we should do this on prod today.

Comment 2

[Robert Helmer [:rhelmer]]

I am taking a prod snapshot before doing this, named "before-bug-1192380"

Comment 3

[Peter Bengtsson [:peterbe]]

We've concluded that the reason crontabber didn't work was because the jobs code hasn't been upgraded from the latest master.

Comment 4

[Robert Helmer [:rhelmer]]

Snapshot is complete, ran truncate on prod:

breakpad=> select truncate_partitions(2);
 truncate_partitions 
---------------------
 t
(1 row)