Files locking might be an issue for scalability?

Marc Powell marc at ena.com
Tue Apr 10 18:53:28 CEST 2007



> -----Original Message-----
> From: nagios-users-bounces at lists.sourceforge.net [mailto:nagios-users-
> bounces at lists.sourceforge.net] On Behalf Of Marc Powell
> Sent: Tuesday, April 10, 2007 11:26 AM
> To: nagios-users at lists.sourceforge.net
> Subject: Re: [Nagios-users] Files locking might be an issue for
> scalability?
> 
> 
> 
> > -----Original Message-----
> > From: Sebastian Ganame [mailto:sganame at yahoo.com.ar]
> > Sent: Tuesday, April 10, 2007 10:48 AM
> > To: Marc Powell; nagios-users at lists.sourceforge.net
> > Subject: RE: [Nagios-users] Files locking might be an issue for
> > scalability?
> >
> > Hi Marc (& others that answered my question),
> >     I share your mindset that if using polling 1K instances every 10
> > seconds (or so) wouldn't make any sense.
> >      When I post file locking might be an issue, I was not thinking
on
> a
> > pull mechanism for services (Nagios trying to retrieve information),
> but
> > in Nagios being pushed by services for this kind of info. As far as
I
> read
> > from documentation, for a push schema, the monitored instance has to
> write
> > a given file inside Nagios structure to update his status.
> >     What do you think about this?
> 
> Passive checks are written to a named pipe that is read by nagios, not
a
> file. Again, file contention isn't going to be an issue afaik as your
OS
> will handle access to the named pipe. You do need to be worried about
> sending results faster than nagios can check them though. Your OS will
> buffer result submissions up to your max pipe size (typically 4096
bytes
> I believe) and you can configure nagios to check external commands as
> often as it possibly can (command_check_interval -1 I believe) but
even
> so, processing 10 results for 1000 host every 10 seconds (1000
> results/second) is not trivial. Again, you'd probably be breaking new
> ground and I'd be surprised if it worked with the current
architecture.
> I submit about 4000 passive results every 5 minutes (~13/second) and
> feel like I have plenty of room to grow but not like that...

Sorry for the followup to my response but I'm just getting up to speed
with what's coming with 3.0. Nagios-3 introduces a new external command
that tells nagios to read further commands from a specified file for
bulk processing. This will eventually be an option for NRPE but it would
be useful in your case.

http://www.nagios.org/developerinfo/externalcommands/commandinfo.php?com
mand_id=131

You would have your external checker write the results to an arbitrary
file in bulk then you would tell nagios to process that file when it was
'full', whatever that criteria was. That would at least get you around
the limitations of the named-pipe processing but you'd still be limited
by how much time was required to process the file and it's commands.

--
Marc

-------------------------------------------------------------------------
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-users mailing list
Nagios-users at 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





More information about the Users mailing list