Symptoms
After upgrading Central Manager, the admin GUI no longer allows you to log in.
The /api/system-info API endpoint indicates that the upgrade is in progress and the system is in maintenance mode.
Fluentd logs stop happening, with the last line indicating that a file is being copied.
nats.fluent_log: {"level":"info","msg":"Initiating platform upgrade"}
nats.fluent_log: {"level":"info","msg":"file copy destination: /vol/local/upgrade/packages/BIG-IP-Next-CentralManager-20.0.2-0.0.68-Update.iso.tmp"}
Lastly, the file at /vol/local/upgrade/packages/ is not the complete file.
Impact
The upgrade fails and the system remains in maintenance mode.
Conditions
-- Upgrading Central Manager
-- The exact conditions that trigger this have not yet been determined.
Workaround
Echo $(kubectl get secret mbiq-db-postgresql -o jsonpath='{.data}' | jq -r '."postgres-password"' | base64 -d)
#Copy the password and supply it into the following command.
?
kubectl exec mbiq-db-postgresql-0 -c postgresql -i -t -- psql -h localhost -U postgres -d bigiq_db -c "UPDATE mbiq_shared.tasks SET state='resumeFromSelfUpgrade', status='running' WHERE task_type='cm-upgrade-task' AND status='running';"
kubectl rollout restart deployment mbiq-upgrade-manager-feature
