Bug with two nagios services starting

[rdedon]

Would this happen to be when 2 users are making changes to the same host/service or just logged in? I want to see if I can try and duplicate this if it is the former and see if it hangs up or what results I may get.

Thanks!

[niebais]

No it would be different hosts. It happened the other day as well for no apparent reason. Nobody was working on the system on Friday and multiple instances of nagios spawned (3 instances). Is it possible that nagios is trying to detect if the Nagios daemon is up or not and it is spawning multiple copies?

[mguthrie]

Although what you're experience "shouldn't" be happening, it appears it is possible on a machine with lots of user activity. There is the /usr/local/nagios/var/nagios.lock file that get's created as soon as the nagios start process is called. I'll have to do some snooping around to see if it looks for the file, or the PID, or what might be happening that could still allow the two instances.

[admin]

There shouldn't be anything in the default XI install that checks to make sure Nagios is running and automatically restarts. This is by design so admins can stop Nagios Core if necessary.

The /etc/init.d/nagios init script could probably use some tweaking to ensure all nagios processes are killed off when the script is called with a "stop" or "restart" option. I'll take a stab at doing this for the next release.

[niebais]

As another follow up on this, the problem occurs occasionally even when only 1 person is applying changes.

[rdedon]

As another follow up on this, the problem occurs occasionally even when only 1 person is applying changes.

now this is very interesting with it ghosting with only 1 user accessing.

[mguthrie]

Our lead developer is going to do some revising on the nagios init script, which should hopefully take care of this in the next release.

[niebais]

We just upgraded to the latest version of Nagios hoping it will solve the issue. However, I discovered something today before we upgraded. We had the dual nagios parent problem. One thing I had to do with the nagios process is a kill -4 on it to make it go away. A regular kill wasn't working so it might have something to do with the signals and it's not handling the kill signal properly.

[mguthrie]

Thanks for the update. I'll forward this along to our lead developer. He's going to try and have the init script revised for the next point release so that this isn't an issue.

[arnab.roy]

Just to add to the pot...I have seen this as well.....but as we are not fully live yet..I havent chased this one up...