Comment 7 for bug 1844456

A similar issue was reported in the dc lab on a sub cloud. The Service group web-services degraded alarm was never cleared. In this case, the lighttpd process was running on the controller-1 but the pid file was not created, as a result, sm determined the process was in a failed state.

2019-10-01T18:53:52.000 controller-1 sm: debug time[358816.750] log<704033> ERROR: sm[22247]: sm_service_fsm.c(1518): Failed to open pid file (/var/run/lighttpd.pid) for service (lighttpd).
2019-10-01T18:53:52.000 controller-1 sm: debug time[358817.326] log<6101> INFO: sm_alarm[22277]: sm_alarm_thread.c(638): Raising alarm (service-group-state) for node (controller-1) domain (controller) entity (web-services).
2019-10-01T18:53:52.000 controller-1 sm: debug time[358817.328] log<6102> INFO: sm_alarm[22277]: sm_alarm_thread.c(1083): Raised alarm (service-group-state) for node (controller-1) domain (controller) entity (web-services), fm_uuid=7fc34e30-6926-450e-9039-10e7e84de61b.

After adding '--make-pidfile' option to start-stop-daemon command in the lighttd initscript, the alarm was eventually cleared.