No data was received from host!

[MichielvM]

Hi all,

I've got this issue with 3 hosts that has been bugging me for some time.
At one of our customers site (300+ monitored hosts) there's 3 Virtual windows 2003 R2 Enterprise x64 hosts that lose connection on a regular (daily) basis at random times.
Error: No data was received from host!
The problem is fixed by logging on to the server and restarting the nsclient service (version: 0.4.1.102).
There's a mix of checks; Nagios built-in (i.e Ping), check_nrpe and check_nt services. Only the check_nt's fail with this error!
I have tried to re-install the client, complete un-install it.

Does anyone have similar experience with check_nt failing at random intervals?

[abrist]

Well, first of all:
Ping is not run through check_nt, it is an active check run completely from the nagios server.
Additionally, nsclient 0.4.x is not necessarily considered stable.

[MichielvM]

I added this info to rule out any network related issues.
The connection remains intact, except for the check_nt checks.

What do you mean by nsclient 0.4.x being not very stable?
What's the alternative? Last I checked at nsclient.org, we're stil in the 0.4.x range.
I will try to install the very latest version and report back.

But besides that, I'd like to point out that I have 30+ Xi servers, monitoring thousands of hosts, and only 3 of them fail like this.
My first bet would not be the client.

[abrist]

Well, we still consider 0.4.x to be a moving target with documentation in constant flux. Due to this, we still suggest 0.3.9. If you have other remote hosts running with the same config on 0.4.x, then I would agree with your assessment. The 0.3.9 alternative is listed as the "stable" agent in the windows wizards and is downloadable from there.

Error: No data was received from host!

1. I doubt check_nt is failing, but to be sure, can you run the check from the nagios server's cli with the verbose (-vvv) flag?
2. Can you check the nsclient logs for any warnings or errors pertaining to this check as well?

[tmcdonald]

In addition, are any other identical Windows 2008 R2 Enterprise working or are these the only three? If these are the only hosts having the problem *and* the only WinServ 2K8R2E hosts, I would lean toward nsclient being the culprit.

[MichielvM]

I had them running the latest (4.105) but that too failed.
This afternoon downgraded to 3.9.x.
We'll see what the next 24 hours will bring.

[slansing]

Great! Keep us apprised!

[MichielvM]

I've had the 24 hour period stretched to span the weekend, and guess what? They're all still up!

Despite the explanation given by Andy that anything above version 3.9 isn't considered stable, I cannot help but wonder why this only happens on 2 out of a few thousand hosts..
The others are all running 4.0/4.1 versions.
Mind you, I will keep this in mind when I run into weird behaviour.

[abrist]

Despite the explanation given by Andy that anything above version 3.9 isn't considered stable, I cannot help but wonder why this only happens on 2 out of a few thousand hosts..

I wish I had a better explanation. Are these windows servers running the exact same platform? Same updates? Were you duplicating the same config across all your servers?
Honestly, the list of things that can cause odd behavior with nsclient is pretty long.

[WillemDH]

You could also check the nsclient.log file in the nsclient installation folder. I've got some useful info from that log in the past. I'm at 0.4.1.101 everywhere. In fact I'm monitoring these nsclient.log filesizes. If one suddenly starts growing, it has helped me in the past. It also alerts me when nsca stops working.
And you can always put a post in Michael Medin's forum: http://www.nsclient.org/forums/

Grtz

Willem