Hey Mark,
I've been stewing on an idea like this as well. I haven't come up with a
perfect solution yet. I know of another user who implemented a large
install and used NAS for the rrdfiles, but I recognize your concerns
there. Would it be plausible to simply mount an additional drive in the
perfdata directory so that all of those writes happen to a separate disk
while still on the local machine?
The other idea I've been thinking about but haven't had time to play
with yet would be to use the performance data processing command to send
the perfdata to the offloaded machine (maybe using xinetd), and then
just drop that data into the perfdata spool so you could have pnp
running on the offloaded machine. From there you could just the web
access for PNP on the 2nd machine. Obviously there are some mechanics to
work out there, and I'm not sure how much bandwidth that would eat up,
but like I said, so far it's just in the idea stage.
On 10/3/2012 9:56 AM, Frost, Mark {BIS} wrote:
Davor,
My concern is more about the actual I/O to the RRD files and not so
much processing the to-be-processed perfdata files (i.e. temporary
files). The heavy I/O is happening on the RRD filesystem and since I
would of course need the RRD files to persist, I would not want to
store them on a ram disk. Plus it would need to be a fairly large ram
disk to hold all the rrd files even if I were willing to lose them all
if a reboot occurred.
We do use ram disks for Nagios status.dat files and spool files (i.e.
things I can afford to lose in a reboot/crash) and it's definitely
been a good thing. It still seems weird to have to do so much
"compensating" for Nagios normal operations for a moderately large
installation (not really even huge) to make it work well. I'm
guessing again that this is going to be vastly improved with Nagios 4
as well. At least no spool files.
Thanks
Mark
*From:*davor grgicevic [mailto:dgrgice...@gmail.com]
*Sent:* Wednesday, October 03, 2012 10:45 AM
*To:* Nagios Users List
*Subject:* Re: [Nagios-users] solutions for off-server PNP4Nagios
perfdata processing?
Hi Mark ...
did you try a using a ram disk
http://exchange.nagios.org/directory/Documentation/Nagios-XI-Documentation/Utilizing-A-RAM-Disk-In-NagiosXI/details
Davor
On Wed, Oct 3, 2012 at 4:33 PM, Frost, Mark {BIS}
<mark.fro...@pepsico.com <mailto:mark.fro...@pepsico.com>> wrote:
Hello. Has anyone come up with solutions for processing Nagios
performance data on a server other than a Nagios server? We've been
processing perfdata results on our Nagios server(s) for a while now
and increasingly it's just eating up too much I/O to make me comfortable.
Yes, we do use rrdcached and yes, I realize that shuffling data around
on different disk spindles and controllers would help, but in today's
world where companies don't like building any kind of physical server
let alone one with all that additional hardware, that's not entirely
an option for us.
I realize that once the perfdata files are on the dedicated graphing
server(s), processing them into RRD files there should be a
no-brainer. My problem is figuring out how to get them there without
say, using a NAS device. (If I/O's a problem locally, I don't want
to shuffle that I/O to an even slower network device).
It would be ideal if somehow there was a process that I could just
send that data to and have it picked up remotely. Like if maybe
Merlin have a special kind of peer that just received a stream of
perfdata or something. Anything else I could imagine would be some
kind of home-grown solution like say pumping events into a messaging
system from the Nagios server(s) and then letting the graphing server
pick them up from the message queue(s). I could also imagine some
kind of fancy-pants module in Nagios 4 that did something like this,
maybe.
Any thoughts would be appreciated.
Thanks
Mark
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Nagios-users mailing list
Nagios-users@lists.sourceforge.net
<mailto:Nagios-users@lists.sourceforge.net>
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when
reporting any issue.
::: Messages without supporting info will risk being sent to /dev/null
--
Davor Grgicevic
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Nagios-users mailing list
Nagios-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting
any issue.
::: Messages without supporting info will risk being sent to /dev/null
--
Mike Guthrie
Technical Team
___
Nagios Enterprises, LLC
Email: mguth...@nagios.com
Web: www.nagios.com
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Nagios-users mailing list
Nagios-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting
any issue.
::: Messages without supporting info will risk being sent to /dev/null