<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii"> <META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7652.24"> <TITLE>NSCA and Latency</TITLE> </HEAD> <BODY>  My Environment: 3 x Dell 2950 Dual DualCore and 8 GB of RAM One system runs checks against our Linux servers One runs checks against our Windows servers We are running SLES10 update 3 Both systems use nsca to send their check results to a third server that displays the service checks for our operators. All three systems are on the same vlan but separate cisco switchs. I am running nsca in daemon mode on the central server with this command /usr/local/nagios/bin/nsca -c /usr/local/nagios/etc/nsca.cfg –daemon Nsca.cfg is as follows: pid_file=/var/run/nsca.pidserver_port=5667#server_address=192.168.1.1nsca_user=nagiosnsca_group=nagios#nsca_chroot=/var/run/nagios/rwdebug=1command_file=/usr/local/nagios/var/rw/nagios.cmdalternate_dump_file=/usr/local/nagios/var/rw/nsca.dumpaggregate_writes=1append_to_file=1max_packet_age=300password=xxxxxxxxxxdecryption_method=14 I just set the aggregate and append options to try and fix the problem they were not set before either way the results are the same. Ok so on the 2 servers doing the checks…. Everything runs fine even with the OCSP running my send_service_check_results script. My script is pretty much straight out of the book. #!/bin/sh# Arguments:# $1 = Hostname of the host (using the $HOSTNAME$ macro)# $2 = Service description of the service (using the $SERVICEDESC$ macro)# $3 = Service status id of the service (using the $SERVICESTATUSID$ macro)# $4 = Output of the Service Check (using the $SERVICEOUTPUT$ macro)/bin/echo "$1","$2","$3","N3 - $4" | /usr/local/nagios/libexec/send_nsca -H 10.10.129.37 -c /usr/local/nagios/etc/send_nsca.cfg -d "," Like I said everything is fine on the 2 servers even with OCSP on. Between the 2 servers we are running about 10k service checks, latency is very low just a few seconds. However if I turn on the NSCA Deamon on the central server my latency creeps up to about 1500+ seconds with in an hour and just gets worse from there on both remotes. The checks that should run every 5 minutes on the 2 remote servers end up running every few hours or less. The central server is doing 0 active checks. I set debug mode and that proved to provide very little insight into the problem. CPU and Mem stats are both very low on all three server. The same thing can be said for the network, network utilization is less than 2% and there are no errors on the interfaces. Overall hardware utilization is 10% or less on these three systems. So my question is has anyone had this kind of problem with NSCA? What am I missing? Should I be batching my service checks on the remote servers? Should I be using xinetd for NSCA instead of deamon mode? Thanks Brady </BODY> </HTML>