perfparse-users

[Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ricardo D. M. <ric...@gm...>]

Hi,

i am currently making some tests with perfparse i've encountered some issues.


On perfparse.sh it stats PERF_FILE="/usr/local/nagios/var/serviceperf.log".
My question is regarding the hostperf.log. What about it? Is this ever
processed?

Another question is regarding the periodically executed script
perfparse.sh. It makes nagios restart. This is annoying. Is this
normal? Is it really necessary?

I think PerfParse is very useful, especially what concerns historical
records and flexibility of showing it.

Many thanks

-- 
Ricardo David Martins

Re: [Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ben C. <Be...@cl...>]

Ricardo,

The hostperf.log is not used in any way.  This program only processes 
the performance data from the plugin, and records it's output and return 
value on each run.

There is an option to delete this file which is present for convenience 
only.

Sorry about the restart.  You can delay PerfParse to every few minute if 
you prefer.  There is a permanent fix for this problem currently in 
development where you will not have to restart Nagios in this way more 
than once a day, and get data within seconds of Nagios publishing it.

I hope you enjoy using it.  Please let us know if there is anything we 
can do, or anything you believe must be included before we get to 
version 1.0.0.

Regards, Ben Clewett.



Ricardo David Martins wrote:

> Hi,
> 
> i am currently making some tests with perfparse i've encountered some issues.
> 
> 
> On perfparse.sh it stats PERF_FILE="/usr/local/nagios/var/serviceperf.log".
> My question is regarding the hostperf.log. What about it? Is this ever
> processed?
> 
> Another question is regarding the periodically executed script
> perfparse.sh. It makes nagios restart. This is annoying. Is this
> normal? Is it really necessary?
> 
> I think PerfParse is very useful, especially what concerns historical
> records and flexibility of showing it.
> 
> Many thanks
> 

Re: [Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ricardo D. M. <ric...@gm...>]

Ben,

I've found something awkward,

on presenting the graphs PP always states "Warning, graph seems to have died".

Is this ok?

Regards

-- 
Ricardo David Martins

On Wed, 08 Sep 2004 14:50:00 +0100, Ben Clewett <be...@cl...> wrote:
> Ricardo,
> 
> The hostperf.log is not used in any way.  This program only processes
> the performance data from the plugin, and records it's output and return
> value on each run.
> 
> There is an option to delete this file which is present for convenience
> only.
> 
> Sorry about the restart.  You can delay PerfParse to every few minute if
> you prefer.  There is a permanent fix for this problem currently in
> development where you will not have to restart Nagios in this way more
> than once a day, and get data within seconds of Nagios publishing it.
> 
> I hope you enjoy using it.  Please let us know if there is anything we
> can do, or anything you believe must be included before we get to
> version 1.0.0.
> 
> Regards, Ben Clewett.
> 
> 
> 
> 
> Ricardo David Martins wrote:
> 
> > Hi,
> >
> > i am currently making some tests with perfparse i've encountered some issues.
> >
> >
> > On perfparse.sh it stats PERF_FILE="/usr/local/nagios/var/serviceperf.log".
> > My question is regarding the hostperf.log. What about it? Is this ever
> > processed?
> >
> > Another question is regarding the periodically executed script
> > perfparse.sh. It makes nagios restart. This is annoying. Is this
> > normal? Is it really necessary?
> >
> > I think PerfParse is very useful, especially what concerns historical
> > records and flexibility of showing it.
> >
> > Many thanks
> >
> 
>

Re: [Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ben C. <Be...@cl...>]

Can you upgrade to version 0.100.4 which should deal with this problem.

Regards, Ben.

Ricardo David Martins wrote:

> Ben,
> 
> I've found something awkward,
> 
> on presenting the graphs PP always states "Warning, graph seems to have died".
> 
> Is this ok?
> 
> Regards
> 

Re: [Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ricardo D. M. <ric...@gm...>]

Ben,

i am concerned with the SIGHUP to nagios process. When nagios shutsdown it
deletes the entire current status and so, it will lose any awaiting
requests. This
will imply nagios to reschedule all checks.=20

Well, wont this badly increase the load?=20

Does this function well with say 500 services on the same server? And
what about 5000?

Do you know of any experience in a production environment with huge loads?

Regards

--=20
Ricardo David Martins
___________________________________________________________________________=
__________________________

AVISO

Esta mensagem (incluindo quaisquer anexos) pode conter informa=E7=E3o
confidencial para uso exclusivo do destinat=E1rio. Se n=E3o for o
destinat=E1rio pretendido, n=E3o dever=E1 usar, distribuir ou copiar este
e-mail. Se recebeu esta mensagem por engano, por favor informe o
emissor e elimine-a imediatamente. Obrigado.

DISCLAIMER

This e-mail (including any attachments) may contain confidential
information for exclusive use of its recipient. If you are not the
intended recipient you must not use, distribute or copy this e-mail.
If you have received this e-mail in error please notify the sender and
delete it immediately. Thank You.

___________________________________________________________________________=
__________________________

[Perfparse-users] Raw history report bug

[Tim W. <tim...@pi...>]

Version 0.100.3 seems to have introduced a bug in the raw history report,
total times (and hence percentages) are no longer calculated:

Output from 0.100.1:

State Count Percent Time 
UP 6153 89.996% 17 days 23:55:32 
WARNING 666 9.741% 1 days 22:45:05 
CRITICAL 18 0.263% 0 days 01:15:48 
UNKNOWN 0 0.000% 0 days 00:00:00 
Start: 2004-08-21 12:08:15   End: 2004-09-10 12:04:42   Span: 19 days
23:56:27  

Output from 0.100.4:

State Count Percent Time 
UP 6153 nan% 0 days 00:00:00 
WARNING 666 nan% 0 days 00:00:00 
CRITICAL 18 nan% 0 days 00:00:00 
UNKNOWN 0 nan% 0 days 00:00:00 
Start: 2004-08-21 12:08:15   End: 2004-09-10 12:04:42   Span: 19 days
23:56:27  

Tim

Re: [Perfparse-users] Raw history report bug

[Ben C. <Be...@cl...>]

Tim,

I re-coded this section.  In version 0.100.1 the stats were only correct 
if there was a complete set of data all with the same time interval.  I 
tried to recode for variable frequency data.

Obviously without much success!

Are you using the 'Gant chart only' option?  Can you let me know the 
filters you are using.  I can get a fix out as soon as I can replicate 
the fix the bug.

Regards,

Ben Clewett.


Tim Wuyts wrote:

> Version 0.100.3 seems to have introduced a bug in the raw history report,
> total times (and hence percentages) are no longer calculated:
> 
> Output from 0.100.1:
> 
> State Count Percent Time 
> UP 6153 89.996% 17 days 23:55:32 
> WARNING 666 9.741% 1 days 22:45:05 
> CRITICAL 18 0.263% 0 days 01:15:48 
> UNKNOWN 0 0.000% 0 days 00:00:00 
> Start: 2004-08-21 12:08:15   End: 2004-09-10 12:04:42   Span: 19 days
> 23:56:27  
> 
> Output from 0.100.4:
> 
> State Count Percent Time 
> UP 6153 nan% 0 days 00:00:00 
> WARNING 666 nan% 0 days 00:00:00 
> CRITICAL 18 nan% 0 days 00:00:00 
> UNKNOWN 0 nan% 0 days 00:00:00 
> Start: 2004-08-21 12:08:15   End: 2004-09-10 12:04:42   Span: 19 days
> 23:56:27  
> 
> Tim
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM. 
> Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Perfparse-users mailing list
> Per...@li...
> https://lists.sourceforge.net/lists/listinfo/perfparse-users
> 

Re: [Perfparse-users] hostperf.log and nagios sighup on perfparse.sh execution

[Ben C. <Be...@cl...>]

Ricardo,

The problem of the SIGHUP will be addressed in the near future.  A 
version of PP which does not use this, and also imports data without 
delay, is in development.

I have no data on how PP runs as far as heavy loading goes.  It would be 
interesting to collect such data...

If you or anybody else can tell us the speed and type of machine you 
have, operating system, memory, CPU.  As well as Nagios data:  Number of 
hosts, services and metrics, and average lines/minute Nagios adds to the 
log file.  Whether you run PP and Nagios on the same machine or not. 
Finally the average load on your server.  I can put any data you return 
to me in a table for general viewing.

There is also CGI data.  Speed of a standard graph (say 2 weeks data) to 
draw, and the size of the total database.

Some people may want to suggest a better form on performance analysis :)

But if you send it, I'll collate it.

Regards, Ben.



Ricardo David Martins wrote:
> Ben,
> 
> i am concerned with the SIGHUP to nagios process. When nagios shutsdown it
> deletes the entire current status and so, it will lose any awaiting
> requests. This
> will imply nagios to reschedule all checks. 
> 
> Well, wont this badly increase the load? 
> 
> Does this function well with say 500 services on the same server? And
> what about 5000?
> 
> Do you know of any experience in a production environment with huge loads?
> 
> Regards
>