After upgrading DX NetOps Performance Management the Data Aggregator dadaemon service fails to start or stay running.
The Data Aggregator is installed using the root user, but is owned and run by a non-root sudo user named dauser.
The status command for the service (systemctl status dadaemon) shows it fails to start and suggests the "journalctl -xe" command be run.
That command shows permissions errors creating or writing to various required files.
-- Unit dadaemon.service has begun starting up.
...
Sep 09 14:16:01 data-agg_host dadaemon[4537]: .lock file found, backing up data & deploy, clearing karaf cache
Sep 09 14:16:01 data-agg_host dadaemon[4537]: Starting IM Data Aggregator
...
Sep 09 14:16:01 data-agg_host dadaemon[4537]: start: Redirecting Karaf output to /opt/IMDataAggregator/apache-karaf-4.2.6/data/karaf.out
...
Sep 09 14:16:01 data-agg_host dadaemon[4537]: /opt/IMDataAggregator/apache-karaf-4.2.6/bin/start: line 95: /opt/IMDataAggregator/apache-karaf-4.2.6/data/karaf.out: Permission denied
Sep 09 14:16:01 data-agg_host dadaemon[4625]: Stopping IM Data Aggregator.
...
Sep 09 14:16:02 data-agg_host dadaemon[4625]: mkdir: cannot create directory '/opt/IMDataAggregator/apache-karaf-4.2.6/data/log': Permission denied
Sep 09 14:16:02 data-agg_host dadaemon[4625]: mkdir: cannot create directory '/opt/IMDataAggregator/apache-karaf-4.2.6/data/tmp': Permission denied
Sep 09 14:16:02 data-agg_host dadaemon[4625]: OpenJDK 64-Bit Server VM warning: Ignoring option UnsyncloadClass; support was removed in 11.0
Sep 09 14:16:02 data-agg_host dadaemon[4625]: /opt/IMDataAggregator/apache-karaf-4.2.6/data/port shutdown port file doesn't exist. The container is not running.
...
Sep 09 14:16:02 data-agg_host dadaemon[4625]: Error stopping the Data Aggregator, error code=3
Sep 09 14:16:02 data-agg_host systemd[1]: dadaemon.service: control process exited, code=exited status=3
Sep 09 14:16:02 data-agg_host systemd[1]: Failed to start Data Aggregator.
When checking ownership the data directory under (default path) /opt/IMDataAggregator/apache-karaf-<version> is owned by root, not the dauser.
When checking the /etc/systemd/system located dadaemon.service and activemq.service files neither has the correct User variable to show the dauser as the service owner.
All supported DX NetOps Performance Management releases
The upgrade was run while the root owned cron job to restart the dadaemon every minute was enabled. During the upgrade it tried to start the dadaemon and broke the installation.
The following steps were used to resolve this.