LINUX: Too many instances: ntpd

[Anoop Sivan]

Hi Guys,
We have an issue with our linux server saying below error
"PSMON: minor CMA(NONE,NONE) LINUX: Too many instances: ntpd"

O/S:Red Hat Enterprise Linux ES release 4 (Nahant Update 8)

# ps -ef | grep ntp | grep -v grep
ntp 5401 1 0 11:42 ? 00:00:00 ntpd -u ntp:ntp -p /var/run/ntpd.pid -g
root 12957 5401 0 11:46 ? 00:00:00 ntpd -u ntp:ntp -p /var/run/ntpd.pid -g

[Matti_Kurkela · Accepted solution]

The ntpd in RHEL 4 is configured to run as a non-privileged user "ntp" for security. This is good, but a non-privileged user cannot adjust the system clock. One possible solution would be for ntpd to create another process to handle the actions requiring root privileges, and run the main NTP server as user "ntp". This may be what your ntpd is doing.

The error message is not a standard RHEL 4 error message. It is probably produced by some monitoring system you have installed to your server. Someone has configured the monitoring system to expect only one instance of ntpd. As there are now two ntpd processes, the monitoring system has followed its configured rules and created an alert message.

Some possible ways to go on:

- If you have other RHEL 4 Update 8 servers running ntpd, does the same happen on all of them? If it happens on all RHEL 4 Update 8 servers with ntpd, consider adjusting the configuration of your monitoring system to allow 1..2 instances of ntpd.

- If it happens only on this server, try killing both ntpd processes and then restarting ntpd. Does it happen again?

- If the issue is reproducible, but only ever happens in this one server, then the system clock of this server might have some issues. Is it a physical server or a virtual machine? Are there any other symptoms of clock problems? (For example, is the system clock very very inaccurate if ntpd is not used?)

- If you have custom SELinux security rules on this server, have you accidentally blocked ntpd from adjusting the system clock?

MK

[Anoop Sivan]

Hello Kurkela,

I appreciate your reply....

1. If i kill one of the ntpd instance, it will respawn after one or two days.

2. We are getting this error for some other RHEL4 boxes also.

Thanks&Regards,
Anoop