Miguel_Tellería
Basic Pleskian
[Plesk 12.0.18 on CentOS 6 with the latest microupdates]
Hi there,
From time time we observe the following phenomenon. Am I the only one?
Plesk nginx front-end (sw-cp-server) is running perfectly and listening on port 8443 but the pidfile (/var/run/sw-cp-server.pid) is missing or corresponds to a non-existing process (probably a previous encarnation).
This creates false watchdog alarms and inhibits any changes in Plesk front-end config (e.g. changing TLS certificate) becase a service sw-cp-service stop doesn't kill the process and the socket is already taken.
The only way to come back from this situation is manually killing the actual sw-cp-server process (easy to find through ps or netstat), actually kill it for a second time (the listener is taken over by another process) and start the process with the service command.
The server has not gone through any high load scenario.
Is there a way to fix this?
Hi there,
From time time we observe the following phenomenon. Am I the only one?
Plesk nginx front-end (sw-cp-server) is running perfectly and listening on port 8443 but the pidfile (/var/run/sw-cp-server.pid) is missing or corresponds to a non-existing process (probably a previous encarnation).
This creates false watchdog alarms and inhibits any changes in Plesk front-end config (e.g. changing TLS certificate) becase a service sw-cp-service stop doesn't kill the process and the socket is already taken.
The only way to come back from this situation is manually killing the actual sw-cp-server process (easy to find through ps or netstat), actually kill it for a second time (the listener is taken over by another process) and start the process with the service command.
The server has not gone through any high load scenario.
Is there a way to fix this?