Peter Klausner | 10 Apr 16:58 2006
Picon

Re: better granularity step 300

> you have to reconfigure whole database definition
>
The smokeping_config page mentions how to define different polling
periods for different probes.
But I don't see how to define different RRA layouts and match them to
the proper probe.
How would that work?

Peter Klausner

Tobias Oetiker | 14 Apr 12:42 2006
Picon
Picon

announce smokeping 2.0.8 released

Hi Users,

I have moved the smokeping website to a new server. It is now
located on

 http://oss.oetiker.ch/smokeping

With this I have also released an update release for smokeping
with the following changes

* adjusted links to new home on oss.oetiker.ch
  -- tobi
* let the name of the dns server be configured in DNS.pm
  -- ark <yesv *ring.wtwf.com>
* make buggy_perl_version test work again
  -- niko, noticed by Fran Boon <flavour *partyvibe.com>

cheers
tobi
--

-- 
ETH Zurich
Tobias Oetiker, IT Support Group D-ITET (ISG.EE)
ETL F24.2, Physikstrasse 3, 8092 Zurich, Switzerland
Phone +41 44 63 25286,  http://people.ee.ethz.ch/oetiker/

Niko Tyni | 15 Apr 10:53 2006
Picon
Picon

Re: better granularity step 300

On Mon, Apr 10, 2006 at 04:58:21PM +0200, Peter Klausner wrote:
> > you have to reconfigure whole database definition
> >
> The smokeping_config page mentions how to define different polling
> periods for different probes.
> But I don't see how to define different RRA layouts and match them to
> the proper probe.
> How would that work?

Hi,

unfortunately the RRA layouts are global at the moment. There's no way to
have different layouts for different probes. This is on the TODO list,
but it's not a very high priority (as there hasn't been much demand for
it so far).

A workaround would be to run separate smokeping processes with different
parameters.

Cheers,
--

-- 
Niko

Niko Tyni | 15 Apr 10:59 2006
Picon
Picon

Re: Error 500 running smokeping.cgi

On Fri, Mar 24, 2006 at 10:35:59AM -0300, Edgar Amorim wrote:

> Sorry if my question is too much basic, but I've ran through the list and
> the similar info didn't match my case. I've got 500 Internal server error
> and the following messages in the /etc/httpd/logs/error_log file.

> The smokeping.cgi has first few lines like:
> #!/usr/bin/speedy -w
> # -*-perl-*-
> 
> use lib qw(/usr/local/rrdtool-1.2.12/lib/perl);

> [Thu Mar 23 19:05:56 2006] [error] [client 127.0.0.1] [Thu Mar 23 19:05:56
> 2006] smokeping.cgi: DynaLoader object version 1.04 does not match
> $DynaLoader::VERSION 1.05 at
> /usr/lib/perl5/5.8.5/i386-linux-thread-multi/DynaLoader.pm line 106. 

Hi,

this means that your rrdtool modules are compiled against a different 
version of perl than your SpeedyCGI. 

See http://lists.ee.ethz.ch/smokeping-users/msg01001.html

HTH,
--

-- 
Niko

Eric Brander | 17 Apr 22:35 2006

Suppress notification of failure for downstream devices if upstream device is down

All,

Is there a way to set alerts and/or notification so that downstream
devices that are monitored do not trigger an alarm if an upstream device
is not responding?

For instance, I monitor sites that are connected via a router and also
the devices beyond that router. If I can't ping the router then I know I
won't be able to reach the devices on the other side either. Right now,
if said router fails, I get my email box full of notifications. I would
like to get notified that an upstream device failed and that is all.

However, I don't want to exclude the downstream devices from
notification if indeed they are down when the upstream device is not.

Is anything like this available in Smokeping?

Also, is there a quick way to see just currently down/not responding
devices at-a-glance without having to go through each menu item?

Thanks,

Eric Brander

amedee | 21 Apr 15:08 2006
Picon

FPing: Use of uninitialized value in division

Hello,

I installed smokeping a few months ago but haven't been able to get in to
work properly.
Running it in debug gives me the following output:

/usr/sbin/smokeping --debug
### Compiling alert detector pattern 'startloss'
### ==S,>0%,>0%,>0%
sub {
    my $d = shift;
    my $y = $d->{loss};
    for(1){
        next if scalar  <at> $y < 4 ;
        next unless defined $y->[-4] and $y->[-4] eq 'S';
        next unless defined $y->[-3]
                        and $y->[-3] =~ /^\d/
                        and $y->[-3] > 0;
        next unless defined $y->[-2]
                        and $y->[-2] =~ /^\d/
                        and $y->[-2] > 0;
        next unless defined $y->[-1]
                        and $y->[-1] =~ /^\d/
                        and $y->[-1] > 0;
        return 1;
    }
    return 0;
}

### Compiling alert detector pattern 'someloss'
(Continue reading)

Peter Klausner | 21 Apr 17:53 2006
Picon

Alert via program fails on raising edge

If you setup an external program for an edgetrigger alert,
the notification fails for the raising edge like so:
---8<----
Apr 20 17:35:10 probe1 smokeping[12626]: Alert someloss was raised for
Test.Localhost
Apr 20 17:35:10 probe1 smokeping[12626]: Use of uninitialized value in
system at /opt/smokeping/lib/Smokeping.pm line 1140.
Apr 20 17:35:10 probe1 smokeping[12626]: Can't exec "": No such file
or directory at /opt/smokeping/lib/Smokeping.pm line 1140.
--->8----

The external progrom _does_ run on falling edges,
but it never sees the promised sixth argument.
The offending code around line 1140 is:

----8<---
    my $what;
    if ($edgetrigger and $prevmatch != $match) {
        $what = ($prevmatch == 0 ? "was raised" : "was cleared");
    }
    if (not $edgetrigger and $match) {
        $what = "is active";
    }
    if ($what) {
    # send something
    ...
    if ( $addr =~ /^\|(.+)/) {
        if ($edgetrigger) {
1140:       system $1,$_,$line,$loss,$rtt,$tree->{host}, ($what =~/raise/);
        } else {
(Continue reading)

Troy | 22 Apr 15:16 2006
Picon

Use of unitialized value in concatenation (.) or string

I've been running smokeping for a number of years and haven't had any
problems.  I'm not sure what triggered this latest problem because I just
started to notice the following line in my logs when I go to my site. The
site is still working fine but every time I go to any page it puts this
line in my logs.

smokeping.cgi [client x.x.x.x]: Use of uninitialized value in
concatenation (.) or string at (eval 23) line 7.

I looked at the line 7 and it states:

use Smokeping 2.000008;

On line 23 of the smokeping.cgi it's just part of the the head1
description.  Can anyone give me an idea of where to start to track down
this problem?

Thanks,
-Troy

Niko Tyni | 24 Apr 14:09 2006
Picon
Picon

Re: Alert via program fails on raising edge

On Fri, Apr 21, 2006 at 05:53:55PM +0200, Peter Klausner wrote:

>     if ( $addr =~ /^\|(.+)/) {
>         if ($edgetrigger) {
> 1140:       system $1,$_,$line,$loss,$rtt,$tree->{host}, ($what =~/raise/);

> It seems this works only, if the arguments to system() are processed
> left to right.
> Otherwise the program, i.e. $1 pulled from the address, gets overwritten
> from the 1st subpattern in ($what =~/raise/) - which is empty.

Thanks! Like Damian Conway says in 'Perl Best Practices': 
 Always give captured substrings proper names.

I'll fix this for the next release.

Cheers,
--

-- 
Niko

Niko Tyni | 24 Apr 14:32 2006
Picon
Picon

Re: FPing: Use of uninitialized value in division

On Fri, Apr 21, 2006 at 03:08:23PM +0200, amedee <at> amedee.be wrote:

> I installed smokeping a few months ago but haven't been able to get in to
> work properly.
> Running it in debug gives me the following output:
> 
> /usr/sbin/smokeping --debug
> ### Compiling alert detector pattern 'startloss'

[...]

This is strange. Isn't there anything like

### fping seems to report in ...
or
### assuming you are using an fping copy reporting in milliseconds

?

Does '/usr/sbin/fping -C 1 localhost' work?

Which version of smokeping and fping do you have?

% dpkg -l smokeping fping

Cheers,
--

-- 
Niko

(Continue reading)


Gmane