[x] Note the number of jobs processed today on the sidekiq dashboard (10)
[x] Enqueue some jobs that won't actually do anything (pulmap will simply ack unknown event types) (I enqueued this 5 times):
bundle exec cap staging rails:console
IndexJob.perform_later({"event" => "UNKNOWNEVENT"}) # do this a bunch of times
[x] pull princeton_ansible main, then run the app's usual ansible playbook, then run the upgrade_redis playbook.
[x] deploy the sidekiq upgrade branch. This will also restart workers
[x] check the number of jobs processed; it should go up otherwise we've lost redis data (yep -- went up to 15)
Troubleshooting: You can use the full url from one box to check the redis connection to the other box, e.g. redis-cli -u redis://dpul-staging1.princeton.edu:6379/0 INFO
Prod:
[x] turn off workers for the app
[x] pull princeton_ansible main, then run the app's usual ansible playbook, then run the upgrade_redis playbook.
hackartisan
commented
12 months ago
We run Redis locally for pulmap
Steps for staging:
Troubleshooting: You can use the full url from one box to check the redis connection to the other box, e.g.
redis-cli -u redis://dpul-staging1.princeton.edu:6379/0 INFOProd: