In VMware Identity Manager (vIDM) or Workspace ONE Access environments, the horizon-workspace service may report a "failed" state within the operating system. Despite this service state, the node's overall health remains optimal, and no loss of vIDM functionality is observed by end-users or administrators.
Executing the command journalctl -u horizon-workspace -n 50 --no-pager on the affected node yields the following error sequence:
Failed to start LSB: IDM tomcat server instance..
<affected_vidm_node_fqdn> su [XXXXX] : pam_unix (su:session) : session opened for user horizon (uid=XXXX) by (uid=0)
<affected_vidm_node_fqdn> su [XXXXX] : pam_unix (su:session) : session closed for user horizon
<affected_vidm_node_fqdn> systemd[1]: horizon-workspace.service: Control process exited, code=exited, status=1/FAILURE
<affected_vidm_node_fqdn> systemd[1] : horizon-workspace.service: Failed with result 'exit-code'.
<affected_vidm_node_fqdn> systemd[1]: Failed to start LSB: IDM tomcat server instance
VMware Identity Manager 3.3.7
This issue is caused by a Process State Synchronization Failure between systemd and the legacy SysV init script management on the vIDM appliance. Consequently, Photon OS is unable to accurately detect that the horizon-workspace service is already running.
Resolution Steps:
1) SSH as "root" to the affected node.
2) Stop the horizon service utilizing the SysV init script command: /etc/rc.d/init.d/horizon-workspace stop
3) Start the horizon service utilizing the systemd service manager: systemctl start horizon-workspace.service
4) Validate the synchronized service status: service horizon-workspace status 5) It shows in "Active" state.