1139465 - decom betafarm.mozillalabs celery

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: WebOps: Labs, task)

Description

[:Atoll]

Per :pmac in bug 739362#c10, betafarm.mozillalabs no longer uses celery, so we can probably decom the 'mozillalabs' bits from generic-celery1 and/or wherever else it is.

Comment 1

[C. Liang [:cyliang]]

I removed the celery configuration pieces in puppet and manually deleted the associated mozillalabs rabbit user manually. [1]   Usually, the subsequent puppet run would shut down the celery processes.

There were two sets of mozillalabs celery processes running on generic-celery1.webapp.phx1.  After puppet ran, one set correctly shut down; the other set did not.  These had to be killed manually. [2]  CPU load on the server looks much more in line with the other generic celery servers.

There appear to have been multiple mozillalabs users on the rabbit cluster in PHX1 (mozillalabs-dev, mozillalabs-stage, and mozillalabs-prod).  After verifying that there were no active queues / channels associated with those users, they were removed from the rabbit configuration. [3]


[1] svn revisions - dev: 101513; stage: 101517; prod: 101521

[2] apache    3100     1  1 Jan30 ?        10:57:11 /usr/bin/python ./manage.py celeryd --loglevel=INFO -f /var/log/celeryd-mozillalabs.log -c 4
apache    3101     1  1 Jan30 ?        10:56:31 /usr/bin/python ./manage.py celeryd --loglevel=INFO -f /var/log/celeryd-mozillalabs.log -c 4
apache    3102     1  1 Jan30 ?        10:57:07 /usr/bin/python ./manage.py celeryd --loglevel=INFO -f /var/log/celeryd-mozillalabs.log -c 4
apache    3103     1  1 Jan30 ?        10:56:40 /usr/bin/python ./manage.py celeryd --loglevel=INFO -f /var/log/celeryd-mozillalabs.log -c 4

[3] svn revision 101527