Operating System - HP-UX
1832935 Members
2952 Online
110048 Solutions
New Discussion

failure to start nagios nrpe in itanium

 
Tingli
Esteemed Contributor

failure to start nagios nrpe in itanium

I tried to install nagios plugins and nrpe to hp-ux systems. I don't have much trouble with hp-ux 11.11. I can not compile the plugins in pa-risc 11.23, but I can use those compiled in 11.11 on them without any problem.

Finally, I tried to install them to itanium virtual systems. Both plugins and nrpe are compiled without any problem. But when I start nrpe daemon, I got the error:
# /usr/local/nagios/nrpe -c /usr/local/nagios/nrpe.cfg -d
/usr/lib/hpux32/dld.so: Unsatisfied data symbol 'rfc931_timeout' in load module '/usr/lib/hpux32/libwrap.so'.
Killed

I compiled them in gcc and cc, but both of them got the same messages.
10 REPLIES 10
Russell Carver
Advisor

Re: failure to start nagios nrpe in itanium

I see one problem

PA-RISC = 11.11 Only
Itanium = 11.23 Only
Both above = 11.3x

Hope that helps
Marcel Burggraeve
Trusted Contributor

Re: failure to start nagios nrpe in itanium

Russell,

your information is outdated.
11.23 has been available for both PA-RISC and Itanium for quite some time now.
Colin Topliss
Esteemed Contributor

Re: failure to start nagios nrpe in itanium

Hi,
I get the feeling you compiled nrpe on one system, then tried to install it on another?

I think you'll find you may have a missing library. On the system where you compiled nrpe, do an ldd on the compiled executable. Do the same on the machine where you are seeing the problem. I think (if I've understood you correctly) you will see a missing library error which will point you in the right direction (you'll need to compile the app that provides you with the missing library).

Your ldd output should look something like the following (apologies, its actually from a Linux system, but it will give you an idea what to look for):

javelin:/proc # ldd /usr/bin/nrpe
linux-gate.so.1 => (0xffffe000)
libssl.so.0.9.7 => /usr/lib/libssl.so.0.9.7 (0x4002c000)
libcrypto.so.0.9.7 => /usr/lib/libcrypto.so.0.9.7 (0x4005d000)
libnsl.so.1 => /lib/libnsl.so.1 (0x40157000)
libc.so.6 => /lib/tls/libc.so.6 (0x4016d000)
libdl.so.2 => /lib/libdl.so.2 (0x4028c000)
/lib/ld-linux.so.2 (0x40000000)

The only other thing it might be is that you compiled against a different version of software that does not have the rfc931_timeout symbol defined anywhere.

Regards

Colin
Tingli
Esteemed Contributor

Re: failure to start nagios nrpe in itanium

Actually, I compiled and ran nrpe in the same system. I have found a few have posted similar error report like mines from the web, but I haven't found any solution for it.

Thanks.
Tom Kunkel
Frequent Advisor

Re: failure to start nagios nrpe in itanium

Hm

Same problem. Downloaded sources and compiled on 11.23 ia64.

No luck.

Latest update for NRPE on HP-UX from Nagios Exchange dates back to 2004 and is for PA-RISC.

Anyone know different?

I guess the alternative is to use some ssh script. Not as elegant but will work.
Dennis Handly
Acclaimed Contributor

Re: failure to start nagios nrpe in itanium

Your problem with the rfc931_timeout unsat is due to a broken libwrap.so. See:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1179523
Tingli
Esteemed Contributor

Re: failure to start nagios nrpe in itanium

I tried to set the rfc931_timeout everywhere, whether in nrpe config file or somewhere else. But none of them work. Finally, I had to write a small script to do the job.
Joseph Harnish
New Member

Re: failure to start nagios nrpe in itanium

Would you mind posting the script you said that fixed your issue? And Possibly what you did?

Thanks

--Joe
Olivier Masse
Honored Contributor

Re: failure to start nagios nrpe in itanium

FWIW, I documented this a few weeks ago and you can check my build notes for nagios/nrpe here:

http://www.mayoxide.com/naghpux/build_notes.txt

Specifics for NRPE are at the end.

Tom Kunkel
Frequent Advisor

Re: failure to start nagios nrpe in itanium

Actually - i ditched the script and NRPE totally (too much work + overhead to push out to all the servers) and now run the nagios server as a SNMP trap host.