Hi,
I wonder if someone ran into this as well.
We have zabbix_agentd 1.8.2 running on Solaris 10 x86_64 and controlled by SMF
Whenever a restart of the agent is executed by cfengine3 we notice that the shared memory segment zabbix was using is not cleaned up. The SMF method uses a pkill -TERM on zabbix_agentd, sleeps for 3 seconds and then starts it
So far we can't reproduce it by running the method by hand. Or instructing cfengine to run the same command which makes it all a bit puzzling.
Did someone else notice zabbix_agentd on Solaris 10 leaving shared memory segments behind? If so, under which conditions did it occur?
I wonder if someone ran into this as well.
We have zabbix_agentd 1.8.2 running on Solaris 10 x86_64 and controlled by SMF
Whenever a restart of the agent is executed by cfengine3 we notice that the shared memory segment zabbix was using is not cleaned up. The SMF method uses a pkill -TERM on zabbix_agentd, sleeps for 3 seconds and then starts it
So far we can't reproduce it by running the method by hand. Or instructing cfengine to run the same command which makes it all a bit puzzling.
Did someone else notice zabbix_agentd on Solaris 10 leaving shared memory segments behind? If so, under which conditions did it occur?
Comment