Description of problem: Pull-provider job invocations are occasionally left in pending state. I cannot reproduce this consistently though. The job task is pending on Actions::ProxyAction Version-Release number of selected component (if applicable): Satellite 6.12 snap 4 How reproducible: Not consistently Steps to Reproduce: 1. have an internal capsule with --foreman-proxy-plugin-remote-execution-script-mode=pull-mqtt 2. have a host with yggdrasil client enabled 3. run jobs against host Actual results: Most jobs execute successfully, occasionally a hanging job occurs Additional info: Found when installing and removing katello-agent on hosts along with running yggdrasil, though not certain if it has a connection to the issue
Could you try running with debug logging enabled on the capsule and satellite, try reproducing it and then attaching production.log and proxy.log from the capsule and satellite and output of journalctl -u yggdrasild from the remote host?
And maybe output of journalctl -u mosquitto from satellite/capsule
Have not found a cause for the rapid reconnects, but the main issue here was caused by kptm not restarting yggdrasil after changing its config.
All the related PRs were merged. The one in Foreman[1] will need to be cherry-picked, katello-pull-transport-migrate could probably be rebased, if we manage to find someone who can actually release it. [1] - https://github.com/theforeman/foreman/pull/9412
Right, that didn't take long. kptm-1.0.2 including the fix is out.
Verified on Satellite 6.12 sn 13, after re-registration yggdrasil is correctly restarted and the rex job works as expected. Also verified that re-registration can cope with a stopped yggdrasil service and with uninstalled yggdrasil package
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: Satellite 6.12 Release), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2022:8506