The error you get from sge_execd
points to that the sge_execd
process is running already and has allocated port 6445
already and therefore a second instance cannot use the port again.
Since sge_execd
and sge_qmaster
are services, they shouldn't be started directly on the commandline but rather through the init-scripts that are shipped with the packages. Additionally, the init-scripts will pick up some SGE related environment variables to make the services work correctly.
From Ubuntu 16.04 and onwards, systemd
is normally used to manage startup scripts. In case of the gridengine
scripts, the systemd
way seems not to work. You should enable the SGE parts to start automatically on boot as follows.
$ sudo update-rc.d gridengine-master enable$ sudo update-rc.d gridengine-exec enable
To start the services correctly for now, you could first stop the instances and the start them through systemd
.
$ sudo killall sge_qmaster$ sudo killall sge_execd
Then check if the processes went away.
$ sudo ps aux | grep sge | grep -v grep
After that try to start the services.
$ sudo systemctl restart gridengine-master.service$ sudo systemctl restart gridengine-exec.service