Page 1 of 2
1 host socket-time-out sometimes
Posted: Tue Aug 05, 2014 1:10 pm
by JohnFLi
I have 1 host machine(well several, but 1 that is annoying me). Windows 2008 Standard. NSCLient version 0.1.1.105
I have 9 checks on it (including host alive) all checks for this host I have set with -t 60 through out the day, it will say socket timeout (for one check or the other). Usually right after, it recovers. I have been logged onto the machine while it's doing this, and it isn't busy. The system has 2 quad-core cpus, and 10 gigs of ram.
Windows firewall is off, and it is on the same vlan as the nagios server. Like I said, it will recover for a bit, then report a socket time out again.....so it isn't a port issue.
Any ideas why this is happeneing?
Re: 1 host socket-time-out sometimes
Posted: Wed Aug 06, 2014 10:01 am
by tmcdonald
What is the actual check you are running?
Re: 1 host socket-time-out sometimes
Posted: Wed Aug 06, 2014 11:20 am
by JohnFLi
it doesn't matter..... disk space, wmi, cpu, memory
Code: Select all
check_command check_nt!CPULOAD!-l 5,85,95
check_command check_nt!USEDDISKSPACE!-l c -w 85 -c 90
check_command check_nt!MEMUSE!-w 85 -c 95
it rarely has an issue with more than one at a time. it will have the socket time out for a few min, then recover, time out again, recover...and it will change on which drive its having an issue with, cpu, memory etc. like I said, it's only with this one machine.
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 2:51 am
by Box293
What does netstat report when executed from a command prompt?
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 10:29 am
by JohnFLi
it reports a lot (system in question is a SQL server)
Anythign in particular you are looking for?
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 11:04 am
by abrist
Maybe an interface going bad? Any errors in the event log, specifically pertaining to its network or the nsclient service?
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 11:21 am
by JohnFLi
Nothign meaningfull in the event logs. Nic is fine as this is a vmware vm.
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 11:58 am
by abrist
If the configs are exactly the same as other monitored windows servers, I would guess that this is an issue with the server itself. I seem to recall an MS bug where after a certain amount of uptime (around a year I believe), the tcp/ip stack can get wonky. Have you tried disabling and re-enabling the nic? Rebooting the server?
Re: 1 host socket-time-out sometimes
Posted: Thu Aug 07, 2014 12:38 pm
by JohnFLi
jeepers!!!
I just checked the event log. SYstem has been up for 43793142 seconds. which is 506.8 days. Yeah.....I think this system needs to be rebooted.
Re: 1 host socket-time-out sometimes
Posted: Fri Aug 08, 2014 9:29 am
by tmcdonald
Let us know how things look after a swift kick to the system
