[08:27] <shadeslayer> you write a systemd service called aptly_db_cleanup, which conflicts with the api unit
[08:27] <shadeslayer> ( systemd has a Conflicts key for this )
[08:27] <shadeslayer> this unit will call aptly db cleanup
[08:28] <bshah> but that systemd unit can't block publisher
[08:28] <shadeslayer> and you simply start/stop this service when you login into the aptly service
[08:28] <bshah> ahhh
[08:28] <shadeslayer> bshah: sure, you simply start/stop the systemd unit from jenkins
[08:28] <bshah> I thought it was .timer etc
[08:28] <shadeslayer> and the jenkins job blocks pub jobs
Description
in my studies to chef systemd_unit the aptly_db_cleanup
service, it seems to me the nci aptly is upstart?! Can you verify this?
https://github.com/blue-systems/pangea-kitchen/tree/
767648f2aed2722c75af110c0a8e71a43cb67d87/site-cookbooks/publisher/templates/
default
Copied from email to Rohan, but any help from anyone would be appreciated.
It also looks like db_cleanup is already run on shutdown, startup. Harald mentioned it did not have to be run as often as repo_cleanup. Is this not sufficient?
Nope, needs to block all jobs on Jenkins somehow so that Jenkins doesn't try and publish while aptly is restarting.
Could it possibly be as simple as jenkins job that blocks pub jobs then restarts aptly, then unblocks pub jobs? I am probably not that lucky..
Actually I think all pub jobs will need Block build if certain jobs are running ( restart_aptly_job ) added to ERB file. Then put restart_aptly_job will need a cron option set to (weekly?) set in ERB
@scarlettclark Sounds about right to me. The tricky part imo is probably how one restarts the job on the publisher.
I think I have the systemd bits sorted. Just have question on publisher.xml.erb not matching server side config.xml I have emailed Harald.
This should be all set, due to @bshah fixing my test. thank you! Ben verified that the size on disk shrunk by epic proportions.