Warning: Attempting to execute the command errors...

[ucemike]

I'm glad you've found a work around, may even be worth updating/re-installing postfix... at this stage anything is worth a shot I suppose.

I did just that (removed the rpm and re-installed) but it didn't help unfortunately. I found that the error started on a specific date and will be looking into anything that was installed that date as well as changes I made to the nagios configs. It's odd that previous to then there was no problem. I'll update what I find if it's related to a nagios config.

[tpc]

ucemike, I'm curious if you found a resolution to this issue. We are experiencing a similar issue. The notify-service-by-email will fail for a subset of our hosts when we add a couple of additional services. We've checked everything we could think of; permissions, extra characters in the config, user environments but we can't seem to crack this one. Still getting the Return code of 127 is out of bounds - plugin may be missing when trying to execute /usr/bin/printf. Even changing the command to echo -e fails.

[ucemike]

As I said for us the only way I could resolve the problem is work around the local smtp delivery by writing our own perl script to handle it.

I can post the perl script I used and the commands.conf entry if you do not have the means to create one.

[jashoemaker]

I just ran into a similar issue. I figured out what is causing it though. If you have more than 4000 or so call to a single check command in nagios, some of them start reporting a return code of 127.
My quick fix was to create an additional check command and use that for part of the checks.

I don't know the exact number that breaks nagios but I would guess that it's 4096.

John A Shoemaker

[jsmurphy]

You know what Jashoemaker said just then is actually fairly close to inline with a problem I had last week. On SLES 10/11 I use check_ping for checking if a host was alive, however I've started using CentOS for the most recent server and using check_ping I started getting return code 127. After doing some digging I realised that the server was running out of pipes or forks (I can't remember exactly what it was running out of but it was something like that), the server was incapable of spawning more instances of ping. Changing it to check_icmp resolved the issue as this seems to utilize resources more efficiently which was explained in a forum post google translated from German that I can no longer find.

So if you happen to be running a Redhat based OS it might be worth while paying attention to how many instances of an application are being spawned and trying to work out if it's hitting that limit. I don't think this bares any relevance to the original problem that the author is having though.