Fedora Infrastructure makes use of Koji dist repos, which are generated by tag2distrepo fedmsg consumer running on Bodhi backend whenever new build is tagged. The problem is that dist repos are deleted after certain time, so we need a mechanism for regenerating old dist repos before they are deleted.
I've implemented a quick PoC dist-repo-regen.py script that regenerates dist repos which are either absent or older than configured age. The script is deployed in staging on bodhi-backend01.stg where it appears to be working. After beta freeze I would like to deploy it on one of production Bodhi backends (bodhi-backend02), test it and if it is working - turn it into a cron job ran once a day.
Cron job was just deployed in staging. Production deployment is planned after freeze.
Thanks! Can we get it to only output on error? Or just log to syslog ?
During development/testing of the script I wanted it to send detailed email to myself, but due to a bug in ansible role, a single cron email was unintentionally sent out to apache (sysadmin-main members). But this bug is fixed now and all further email should be sent only to mizdebsk. After development/testing is done I will silence the script and make it emails only errors, to root (sysadmin-main).
:thumbsup:
The cron job is now deployed in production - it should regenerate dist-repos that are either missing or older than 5 days.
Metadata Update from @mizdebsk: - Issue untagged with: unfreeze - Issue priority set to: None (was: Waiting on Assignee)
Metadata Update from @mizdebsk: - Issue close_status updated to: Fixed - Issue priority set to: Waiting on Assignee - Issue status updated to: Closed (was: Open) - Issue tagged with: unfreeze
Login to comment on this ticket.