SSH Check monitor fails, but works manually via CLI

[aylesworth]

Hello,

We have an SSH check that runs a remote script checking the license count for an application. When running the check via the actual service monitor, I get this result (I've changed IP information below):

[nagios@nagiosxi ~]$ /usr/local/nagios/libexec/check_by_ssh -H [IP HERE] -C "~/check_ud.sh"
Remote command execution failed: Permission denied, please try again.

If I log into the server itself, and su - nagios from the root, running the same command, I get the expected output:
[nagios@nagiosxi ~]$ /usr/local/nagios/libexec/check_by_ssh -H [IP HERE] -C "~/check_ud.sh"
Password:
OK - 341 is 71% of 478 Unidata Licenses used.|unidata=71%;90;95

It seems like the keypair/fingerprint is not present, but I've added that both via the IP and the hostname to known_hosts. I've also created a keypair on the nagios server itself and added the public key to the monitoring endpoint server.

What am I missing?

Thanks!

[pbroste]

Hello aylesworth,

Please check the permissions and ownership on your "~/check_ud.sh" script. should be:

Code: Select all

chown nagios:nagios  ~/check_ud.sh

and

Code: Select all

chmod 755 ~/check_ud.sh

Thanks,
Perry

[aylesworth]

Hi Perry,

Thanks for the suggestion, not sure what happened but re-issuing the keypair and re-adding that to the host fixed the issue.

This can be closed!

[gormank]

Have you tried to ssh to the remote host as nagios from the nagios host? If it prompts for a password, something in the ssh setup probably needs fixed.