Nagios Support Forum

Hello everyone,

Yesterday I noticed a problem when I attempted to acknowledge service alerts. During the commit, after entering a comment, i get an error; the commit could not take place and the permissions on /usr/local/nagios/var/rw/nagios.cmd may be incorrect:

[root@sms1 rw]# ll
total 0
prw-rw----. 1 nagios nagcmd 0 Jun 24 17:05 nagios.cmd
srw-rw----. 1 nagios nagcmd 0 Jun 24 17:05 nagios.qh
[root@sms1 rw]# pwd
/usr/local/nagios/var/rw
[root@sms1 rw]#

Nagios XI 2014R2.7

So I change the permissions on nagios.cmd using chmod a+w, the commit works. But then later the permissions on the nagios.cmd named pipe default back to 660 and the commits fail again with the same error.

Is there something re-setting the permissions on that named pipe? What else could be causing the commits to fail?

I was loged in as nagiosadmin when I did these commits. I have not tried as a normal user.

Can you run the following command and show us the output?

This may be the problem. There was no output on the sms1 machine which hosts Nagios XI because all users are mounted using the autofs service.

[root@sms1 home]# pwd
/home
[root@sms1 home]# ls
geneb nagios
[root@sms1 home]# grep nag /etc/group
[root@sms1 home]#


On the machine where we administer users, I did add a nagios user and nagios / nagcmd groups, as follows:

[root@cm1 ~]# grep nag /etc/group
nagios509:nagios
nagcmd510:nagios
[root@cm1 ~]#

Would adding the proper nagios / nagcmd entries in the /etc/group file on sms1 fix the issue? I'm doing this all on the development system so trying things out doesn't break anything critical in my case.

They also need to be part of the apache or whatever group you run HTTPD as: