Satellite: How a stuck task can be removed. Always open a support case. (solution)
On my home lab RH Satellite 6 I had a minor issue at first, Kickstart 7.6 stopped to sync.
This week I decided to drop RHEL 7.6 content, and now the sync issue became a major issue. The task to remove the repository failed due to the old locks.
I opened a support case and recorded the commands performed by the Red Hat Engineer Amar, Huchchanavar who solved the issue.
to avoid support issues: open a support ticket, and ask whether the actions below are save to perform.
At the WebUI
Set dynflow_allow_dangerous_actions
to Yes
(only during the fixing session)
On the CLI
hammer ping
sudo su - postgres -c "psql -d foreman -c 'select label,count(label),state,result from foreman_tasks_tasks where state <> '\''stopped'\'' group by label,state,result ORDER BY label;'"
foreman-maintain service status
foreman-maintain service stop
foreman-maintain service start
sudo su - postgres -c "psql -d foreman -c 'select label,count(label),state,result from foreman_tasks_tasks where state <> '\''stopped'\'' group by label,state,result ORDER BY label;'"
foreman-rake foreman_tasks:cleanup TASK_SEARCH='result == pending' STATES='stopped' VERBOSE=true NOOP=true
foreman-rake foreman_tasks:cleanup TASK_SEARCH='result == pending' STATES='stopped' VERBOSE=true
katello-service restart
hammer ping
At the WebUI
Return the value of dynflow_allow_dangerous_actions
back to No
RJ Hinton has a great addition to avoid other issues
- reposyncs do not have download-policy immediate set:
'YOUR_VALID_ORG_GOES_HERE' is a fake label to make it generic, which you have to change to make it work for your situation.
hammer repository list --organization-label='YOUR_VALID_ORG_GOES_HERE'|\
awk '{print $1}' | egrep -v '\||ID' | \
less | \
perl -ne 'print;chomp;system("hammer repository update --download-policy immediate --id $_")'
*Errors to ignore:*
Could not update the repository:
Validation failed: Download policy Cannot set attribute download_policy
for content type file
- The task history is too long
foreman-rake foreman_tasks:cleanup TASK_SEARCH='label ~ *' AFTER='100s' VERBOSE=true