NSCA using PROCESS_FILE

[Thomas Guyot-Sionnest]

If someone else don't beat me up on time I'll see if I have time for that in
the upcoming days.

However I disagree with the way of doing it. I'd rather have nagios write to
a pipe (We already do this succesfully to pipe performance data into cacti;
a perl daemon holds the data for some time) and have the perl damon do the
batch updates. This would work like this:

- With reaper delay set to 0:
  Block on read, fork and spawn send_nsca/clear the buffer on each loop run.
- With reaper delay set to >0:
  Set an alarm to <reaper delay>, block on read. The alarm handler fork and
spawn send_nsca/clear the buffer.
I'm not sure yet the possible interations of signals and pipe reads, so I
might need non-blocking read there. Our Perl daemon is already non-blocking
as it handle network connections from the cacti script as well.

Thomas

> -----Original Message-----
> From: nagios-devel-bounces at lists.sourceforge.net 
> [mailto:nagios-devel-bounces at lists.sourceforge.net] On Behalf 
> Of Ethan Galstad
> Sent: February 22, 2007 14:34
> To: Gerd Mueller; Nagios Developers List
> Subject: Re: [Nagios-devel] NSCA using PROCESS_FILE
> 
> Any takers (Thomas?) on writing a HOWTO on doing this?  I was 
> going to 
> write up some documentation on doing bulk transfers with NSCA 
> and using 
> the perfdata files for data output.  Since I'm busy with 
> Nagios 3 docs, 
> this will have to wait a week or so.
> 
> If someone wants to write a HOWTO first, I'll ship them an 
> as-yet-unreleased new Nagios t-shirt that is sure to get me 
> and anyone 
> who wears it in trouble with the NSA or the Department of Homeland 
> Security. :-)
> 
> Here's what I was going to writeup:
> 
> 1. Use perfdata file to write passive check information to a 
> file in the 
> format used by send_nsca
> 
> 2. Rotate the perfdata file every 60 seconds
> 
> 3. A cron job runs every minute and pipes the rotated 
> perfdata file to 
> the send_nsca client.
> 
> Bam - bulk data transfer at 60 second intervals.  This should help on 
> both the central and distributed servers.  A better 
> performance increase 
> will happen when the NSCA daemon uses the PROCESS_FILE 
> command when it 
> receives bulk results from the send_nsca clients.
> 
> 
> Gerd Mueller wrote:
> > We are already using ocsp sweeper from nagiosexchange. (at 
> Ton: we need ochp as well) This improves the performce. but 
> we still have latencies upto 100 sec. :( So I would call this 
> just a workaround.
> > 
> > Gerd
> > -- Ursprüngl. Mitteil. --
> > Betreff:	Re: [Nagios-devel] NSCA using PROCESS_FILE
> > Von:	"Thomas Guyot-Sionnest" <Thomas at zango.com>
> > Datum:		22.02.2007 18:54
> > 
> > If I were you I'd use Nagios ability to write performance data to a
> > file/pipe for ohcp/oscp functionality.
> > 
> > Send the data to a pipe and have a Perl (Or anything else 
> you want) daemon
> > read that pipe and spawn send_nsca processes. One good 
> thing with that also
> > is that you could send batch updates since send_nsca can 
> receive multiple
> > results in a single run.
> > 
> > Thomas
> > 
> >> -----Original Message-----
> >> From: nagios-devel-bounces at lists.sourceforge.net 
> >> [mailto:nagios-devel-bounces at lists.sourceforge.net] On Behalf 
> >> Of Gerd Mueller
> >> Sent: February 22, 2007 10:56
> >> To: Nagios Developers List
> >> Subject: Re: [Nagios-devel] NSCA using PROCESS_FILE
> >>
> >>
> >> Am Dienstag, den 20.02.2007, 08:48 -0600 schrieb Ethan Galstad:
> >>> Gerd Mueller wrote:
> >>>> Hi list!
> >>>>
> >>>> Is there a nsca which supports already the Nagios3 
> >> external command
> >>>> PROCESS_FILE? Or any suggestions how to use PROCESS_FILE for
> >>>> inter-Nagios-communication with ohcp/oscp? 
> >>>>
> >>>> Cheers,
> >>>>
> >>>> Gerd  
> >>> NSCA hasn't been modified to use the new PROCESS_FILE 
> >> command, but I'll 
> >>> start working on that in the next few weeks.  Right now I'm 
> >>> concentrating on documentation.
> >> I really need this new nsca feature. Inter-master-slaves 
> communication
> >> in our test environment (about 3000 hosts and livesystem will 
> >> have much
> >> more hosts) via nsca causes the slaves' latency to grow 
> massiv (up to
> >> hours!). Whereas a slave with same hardware and same 
> amount of checks
> >> but without any ocsp/ochp does nearly nothing (latency < 1sec.!). 
> >>
> >> I think the major bottleneck - the serial execution of 
> ocsp/ochp still
> >> exists in Nagios3. So Ton's comment on this
> >> (http://altinity.blogs.com/dotorg/2006/11/caching_nsca_da.html)  is
> >> still true.  
> >>
> >> I expect the new PROCESS_FILE option to avoid this 
> bottleneck. The new
> >> nsca client has to be fast enough without any slowing down 
> the slave
> >> itself while reading and caching. But I really would 
> appreciate if the
> >> main cause of the problem - the ocsp/ohcp executions - could be
> >> deserialized.
> >>
> >> Let me now if I can support the development with testing, 
> coding, ...
> >>
> >> Gerd
> >>
> 
> 
> 
> Ethan Galstad,
> Nagios Developer
> ---
> Email: nagios at nagios.org
> Website: http://www.nagios.org
> 
> --------------------------------------------------------------
> -----------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the 
> chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge
> &CID=DEVDEV
> _______________________________________________
> Nagios-devel mailing list
> Nagios-devel at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nagios-devel
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3076 bytes
Desc: not available
URL: <https://www.monitoring-lists.org/archive/developers/attachments/20070222/25536752/attachment.bin>
-------------- next part --------------
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
-------------- next part --------------
_______________________________________________
Nagios-devel mailing list
Nagios-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-devel