NTP CRITICAL: Offset unknown

[jbennett]

We have implemented an NTP sync check on all of the NRDS checks that we are rolling out right now.

I'm finding that a few locations will report critical in Nagios but the NTP sync appears to be just fine on the host.

Code: Select all

# ntpdc -p
     remote           local            st  poll reach  delay   offset    disp
=======================================================================
=LOCAL(0)        127.0.0.1       10   64  377 0.00000  0.000000 0.03061
*timeserver1     [host ip]          2   64  377 0.00000 11.053003 0.00580

I am seeing a number of people reporting this same issue and that it is a known bug in previous versions and should haev been fixed in 1.4.12?

Code: Select all

# ./check_ntp_time -V
check_ntp_time v2051 (nagios-plugins 1.4.13)

Is this still a known issue with this release?

[abrist]

Can we see the full details and definition of the check? I am curious to what values you have specified in the check and the type of alerts you are receiving.

[jbennett]

This may have just been resolved with an update to the 1.4.16 plug-ins package.

The NTP warnings appear to have gone away for now. If they come up again, I will post.

[slansing]

Great let us know how the update works out for you.

[jbennett]

Looks like it didn't resolve the issue:

When run on the host that's reporting Offset unknown in Nagios:

Code: Select all

# /usr/local/nagios/libexec/check_ntp_time -H localhost -w 0.5 -c 1
NTP OK: Offset -1.147855073e-07 secs|offset=-0.000000s;0.500000;1.000000;

Nagios shows:

Code: Select all

Check NTP Sync	
2m 27s
NTP CRITICAL: Offset unknown

It appears that NRDS has NOT pushed out the newer check routines that I have updated with the 1.4.16 plugins release.

On the Nagios box:

Code: Select all

# /usr/local/nagios/libexec/check_ntp_time -V
check_ntp_time v1.4.16 (nagios-plugins 1.4.16)

On the host:

Code: Select all

# /usr/local/nagios/libexec/check_ntp_time -V
check_ntp_time v2051 (nagios-plugins 1.4.13)

However, when I check the date of the service check on the host, it appears to have been touched today at about the time that I saved the NRDS config on the Nagios server:

Code: Select all

# ls -lh /usr/local/nagios/libexec/check_ntp_time
-rwxr-xr-x 1 nagios nagios 65K Jun 11 11:00 /usr/local/nagios/libexec/check_ntp_time

The nrds.cfg file was updated with the save of the config on the Nagios server, they both show version 8:

Code: Select all

 # ls -lh /usr/local/nrdp/clients/nrds/nrds.cfg
-rwxrwxr-x 1 nagios nagios 4.5K Jun 11 11:00 /usr/local/nrdp/clients/nrds/nrds.cfg

How can I force the updated plug-ins to be updated on the hosts? I'm hoping that this will resolve the issue if it's an actual bug with the check.

[scottwilkerson]

You will need to copy the plugins the the appropriate folder, likely

Code: Select all

/usr/local/nrdp/plugins/Linux/i386/

or

Code: Select all

/usr/local/nrdp/plugins/Linux/x86_64/

see page 3 of the following file on the order NRDS searches for plugins
http://assets.nagios.com/downloads/nagi ... h_NRDS.pdf

NRDS searches for the plugin in the following order (based on the CLIENTS operating system, Arch) and delivers the first match

Code: Select all

/usr/local/nrdp/plugins/$OS$/$ARCH$
/usr/local/nrdp/plugins/$OS$
/usr/local/nrdp/plugins/Generic/$ARCH$
/usr/local/nrdp/plugins/Generic
/usr/local/nagios/libexec

[jbennett]

I just copied the check_ntp* checks from the /usr/local/nagios/libexec/ directory to the /usr/local/nrdp/plugins/Linux/i386/ directory and now I'm getting 'Error code 126 - check plugin'.

On the clients:

Code: Select all

# uname -m
i686

On the Nagios server:

Code: Select all

# uname -m
x86_64

When I run the check from the i386 folder on the Nagios box, I get the following:

Code: Select all

# ./check_ntp_time -H xxx.xxx.xxx.xxx
NTP OK: Offset -26.2263124 secs|offset=-26.226312s;60.000000;120.000000;

When I run it on one of the clients, I get the following:

Code: Select all

:/usr/local/nagios/libexec # ls -lh check_ntp_time
-rwxr-xr-x 1 nagios nagios 141K Jun 12 07:59 check_ntp_time
:/usr/local/nagios/libexec # ./check_ntp_time
-bash: ./check_ntp_time: cannot execute binary file

I'm guessing that since I copied the files from the /libexec folder on my Nagios box into the i386 folder and then those files were pushed to a 32-bit system, that would be my problem.

How can I get the new check to be compiled to work for a 32-bit system?

[abrist]

You will need to rebuild the nagios-plugins package on a 32 bit system, or set "export CFLAGS=-m32" on a 64 bit system before you build the package.

[jbennett]

Thanks.

That took care of the issue with compiling, but I'm still getting the offset unknown error. Everything came up ok initially, now I'mg getting locations with the offset unknown alert.

When I run it from the client command line, I get the error as well.

Code: Select all

# ./check_ntp_time -H localhost
NTP CRITICAL: Offset unknown|
# ./check_ntp_time -V
check_ntp_time v1.4.16 (nagios-plugins 1.4.16)
# ntpdc -p
     remote           local      st poll reach  delay   offset    disp
=======================================================================
=LOCAL(0)        127.0.0.1       10   64   17 0.00000  0.000000 0.96858
*timeserver1     xxx.xxx.xxx.xxx    2   64   17 0.00098  4.956048 0.00580

[abrist]

What do get if you just run:

Code: Select all

ntpdc -p