I have a problem with availability reports (Nagios core 3.2.3).
if I call avail.cgi to generate CSV report for all hosts and services for a given day it uses 100% cpu ,and in most cases eit nds up with a timeout after 10 minutes.
I observe growing memory usage of the avail.cgi process in top; iostat shows pretty no IO activity.
Example call:
Code: Select all
curl -u [user:pass] http://myhost:1080/nagios/cgi-bin/avail.cgi?show_log_entries=&host=all&service=all&timeperiod=custom&smon=08&sday=18&syear=2011&shour=00&smin=00&ssec=00&emon=08&eday=19&eyear=2011&ehour=00&emin=00&esec=00&rpttimeperiod=&assumeinitialstates=yes&assumestateretention=yes&assumestatesduringnotrunning=yes&includesoftstates=no&initialassumedservicestate=0&backtrack=4&csvoutput=
Code: Select all
Checking services...
Checked 325 services.
Checking hosts...
Checked 67 hosts.
Unfortunately avail.cgi doesn't provide CSV output for a single host or hostgroup, same for services;
such reports are only available i nHTMP form which is bad for further processing.
I need to reformat the data, therefore I have to use CSV reports, bu tapparently with zero granularity (all/all) it's not performing well.
No idea how to investigate and fix it, any hints?
BTW is there a port of patch http://www.nagios-portal.org/wbb/index. ... eadID=3315 for newer versions?
I use Nagios 3.2.3, could update to 3.3.1 but can't consider downgrade to older one. If it were possible I'd prefer to generate CSV reports per hostgroup/servicegroup instead of doing this for all/all and filtering the huge output per awk afterwards.
Thanks in advance
P.S. the same worked on another nagios box (single-cpu SuSE 11.1 VMWare image) but doesn't work on the new 2-cpu SuSe 11.1 box, the configuration seems to be the same but...