It looks like we had a successful run after that, so this one is likely transient (but i’m not sure how often it happens). I can raise the priority with the pulp team
I’ve seen it twice in the pipeline and once locally in the past 48 hours. Every time on EL8, only for fresh installs. Perhaps we see this on EL8 more often because it tends to be all around faster allowing the race condition to trigger more? Please do raise the priority as this will get annoying
Additionally we are seeing failures with our bats test not finding foreman-maintain command, doing some debugging I see:
# `which foreman-maintain' failed
# which: no foreman-maintain in (/usr/libexec:/sbin:/bin:/usr/sbin:/usr/bin)
not ok 4 check service status
# (in test file fb-test-katello.bats, line 41)
# `foreman-maintain service status' failed with status 127
# /tmp/bats.62678.src: line 41: foreman-maintain: command not found
However, I can confirm the package exists at the time it is ran and the command exists when I ssh onto the and run it myself:
[root@pipe-katello-server-nightly-centos8-stream vagrant]# which foreman-maintain
/bin/foreman-maintain
[root@pipe-katello-server-nightly-centos8-stream vagrant]# foreman-maintain --help
Usage:
foreman-maintain [OPTIONS] SUBCOMMAND [ARG] ...
Parameters:
SUBCOMMAND subcommand
[ARG] ... subcommand arguments
Subcommands:
health Health related commands
upgrade Upgrade related commands
service Control applicable services
backup Backup server
restore Restore a backup
packages Lock/Unlock package protection, install, update
advanced Advanced tools for server maintenance
content Content related commands
maintenance-mode Control maintenance-mode for application
Options:
-h, --help print help
This seems to happen on CentOS 8 stream locally, and both 8 stream and 8 in our CI system.
Turns out with the change to not enable katello-agent infrastructure by default, we are ensuring that qpid-tools is not present. Since the katello RPM relies on qpid-tools this caused yum to remove qpid-tools and thus katello and thus rubygem-foreman_maintain. This PR should fix this: