proc.num sending back 0

[hm55]

Hello,

When i start the ntpd process on my centOS distrib and check if zabbix can monitor him with :

# zabbix_agent -t proc.num[ntpd]

proc.num[ntpd] [u|1]

i have 1 and when i look in log file with debug mode i have 0

8259:20150803:161543.121 Requested [proc.num[ntpd]]
8259:20150803:161543.127 Sending back [0]

Do you know why i can't get the real value ?

Thanks !

[akbar415]

Hello,

When i start the ntpd process on my centOS distrib and check if zabbix can monitor him with :

i have 1 and when i look in log file with debug mode i have 0


Do you know why i can't get the real value ?

Thanks !

On Zabbix server. Try this

zabbix_get -s hostname -k"proc.num[ntpd]"

The result must be "1"

[hm55]

It's same thing (0) and in log i see the same Sending Back

[akbar415]

It's same thing (0) and in log i see the same Sending Back

ps -ef |grep ntp |grep -v grep
ntp       1471     1  0 10:50 ?        00:00:00 [B][COLOR="green"][FONT="Arial Black"]ntpd[/FONT][/COLOR][/B] -u ntp:ntp -p /var/run/ntpd.pid -g

What's is in green is what you need to parse as argument to zabbix.

[hm55]

And it's preciselly what i have in ps command but always sending back 0 in zabbix_get and in debug log

#ps -ef |grep ntp |grep -v grep
ntp 8160 1 0 Aug03 ? 00:00:00 ntpd -u ntp:ntp -p /var/run/ntpd.pid -g

[hm55]

Same problem with nginx also

[hm55]

Up !
(CentOS 6.5 | Zabbix 2.4.6)

[akbar415]

zabbix_agent -s hostname -k"system.uptime"
or
zabbix_agent -s hostname -k"agent.version"

To check if the communication between agent and server are ok.

[kloczek]

And it's preciselly what i have in ps command but always sending back 0 in zabbix_get and in debug log

Zabbix agent is not running from root account.
Probably you have enabled SELinux and/or you have blocked access to procfs for zabbiix user.

[hm55]

Zabbix agent is not running from root account.
Probably you have enabled SELinux and/or you have blocked access to procfs for zabbiix user.

Ok,
So the selinux-policy was in version 3.7.19-231 that have a policy mistake when i update the policy it fix the problem !

Thanks !