Larry Low | 1 Feb 02:02 2008

Master/Slave Setup - Master No Polling

I am getting some weird behavior with my Master/Slave setup.  The Master is
set not to poll.  I randomly get the empty graphs for the master.  When I
only see the Slave graphs I am unable to zoom in and the charts show the
numbers but no graphs.  I don't get anything in the http error_log.

Any ideas?

Also line 631 in Smokeping.pm has issues where hide is not defined.  I went
ahead and added a check of the variable before checking its value to get rid
of the errors filling up my error_log.

----
Larry Low
4150 N Drinkwater Blvd., 5th Floor
Scottsdale, AZ  85251
Office: 480.385.7045
E-mail: llow <at> telesphere.com'

Telesphere Networks, Inc

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users

Rui Meireles | 1 Feb 12:48 2008
Picon

CheckLoss - Bug or Feature?

Hi. I’m trying Smokeping for the first time.

I don’t understand the mechanics of CheckLoss matcher.

 

CheckLoss(l=>3,x=>4)

 

Does this mean an alarm will be raised if there are:

1 - at least 3 packets loss in the last 4 pollings?

2 - at least 3 packets loss in each of the last 4 pollings?

3 - 3% packet loss in the last 4 pollings?

4 - 3% packet loss in each of the last 4 pollings?

 

Despite this, this matcher is not working well. When a device becomes unreachable, usually I start receiving flapping alarms, that go on (Raised, Cleared, Raised, Cleared, …) until there are no 100% losses in the small historic that comes with the e-mail.

When the device comes back to life, this usually happens too.

 

 

Ex.:

#####################################################

Thu Jan 31 20:03:43 2008

 

Alert "someloss" was raised for http://172.28.154.41/cgi-bin/smokeping/smokeping.cgi?target=WAN.Rede_Tecnica.Rui_Meireles

 

Pattern

-------

CheckLoss(l=>3,x=>4)

 

Data (old --> now)

------------------

loss: 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 100%, 100%, 100%, 100%, 100%, 100%, 100%, 100%

rtt: 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, U, U, U, U, U, U, U, U

 

Comment

-------

Perdas Detectadas

 

#####################################################

Thu Jan 31 20:08:43 2008

 

Alert "someloss" was cleared for http://172.28.154.41/cgi-bin/smokeping/smokeping.cgi?target=WAN.Rede_Tecnica.Rui_Meireles

 

Pattern

-------

CheckLoss(l=>3,x=>4)

 

Data (old --> now)

------------------

loss: 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 0%, 100%, 100%, 100%, 100%, 100%, 100%, 100%, 100%, 100%

rtt: 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, 0ms, U, U, U, U, U, U, U, U, U

 

Comment

-------

Perdas Detectadas

 

#####################################################

 

 

In the example above (2 of the many e-mails I received) why was the alarm cleared? There is no reason for that!

And why is it showing such a big historic, if I specified 4 values?

 

If anyone could help me, I would deeply appreciate it. All I want is to generate alarms when there are some (few) packet losses, and terminate them as soon as the situation resolves.

 

Thanks.

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
Aaron Lewis | 1 Feb 19:57 2008
Picon

how low can it go?

Before I dig into setting up smokeping, I would like to ask if I can
monitor very small latency numbers--sub millisecond?  All nodes are in
a single building LAN.  I need to monitor latency over time to track a
baseline, compare changes, etc.  Everything I've looked at on Windows
(What's Up Gold, IPCheck) won't measure smaller than 1 ms.  Is this
because the built in PING in windows doesn't resolve lower than 1 ms?
I notice in any Linux distro, PING gives me numbers like 0.255 ms
which is great.  So will smokeping work at this level?

If so, what is the easiest way for someone to get smokeping running?
I can setup a dedicated machine, maybe a few years old (P4, 1GB RAM,
120+ hdd) polling about 100 nodes on the LAN. I would probably want to
monitor our ISP's DNS, web, and mail servers, which would be measured
on the more "normal" ms level. Do some distros have smokeping
installed already?  I'm fairly new to Linux, but work as a Windows
system admin, Net+ certified.  I'm familiar with basic CLI
administering though SSH.  Is there something like Remote Desktop for
Linux?  Thanks!

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users

Ralf Hildebrandt | 1 Feb 20:24 2008
Picon

Re: how low can it go?

* Aaron Lewis <pcrequest <at> gmail.com>:
> Before I dig into setting up smokeping, I would like to ask if I can
> monitor very small latency numbers--sub millisecond?

Yes.

--

-- 
Ralf Hildebrandt (i.A. des IT-Zentrums)         Ralf.Hildebrandt <at> charite.de
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                    send no mail to plonk <at> charite.de

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
Ralf Hildebrandt | 1 Feb 20:24 2008
Picon

Re: how low can it go?

* Aaron Lewis <pcrequest <at> gmail.com>:

> If so, what is the easiest way for someone to get smokeping running?

apt-get install smokeping

--

-- 
Ralf Hildebrandt (i.A. des IT-Zentrums)         Ralf.Hildebrandt <at> charite.de
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                    send no mail to plonk <at> charite.de

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
Vinny Abello | 1 Feb 20:25 2008

Re: how low can it go?

I have smokeping monitoring several servers on the same LAN segment and it reports in the range of about
180µ to 500µ or microseconds.

Aaron Lewis wrote:
> Before I dig into setting up smokeping, I would like to ask if I can
> monitor very small latency numbers--sub millisecond?  All nodes are in
> a single building LAN.  I need to monitor latency over time to track a
> baseline, compare changes, etc.  Everything I've looked at on Windows
> (What's Up Gold, IPCheck) won't measure smaller than 1 ms.  Is this
> because the built in PING in windows doesn't resolve lower than 1 ms?
> I notice in any Linux distro, PING gives me numbers like 0.255 ms
> which is great.  So will smokeping work at this level?
> 
> If so, what is the easiest way for someone to get smokeping running?
> I can setup a dedicated machine, maybe a few years old (P4, 1GB RAM,
> 120+ hdd) polling about 100 nodes on the LAN. I would probably want to
> monitor our ISP's DNS, web, and mail servers, which would be measured
> on the more "normal" ms level. Do some distros have smokeping
> installed already?  I'm fairly new to Linux, but work as a Windows
> system admin, Net+ certified.  I'm familiar with basic CLI
> administering though SSH.  Is there something like Remote Desktop for
> Linux?  Thanks!
> 
> _______________________________________________
> smokeping-users mailing list
> smokeping-users <at> lists.oetiker.ch
> https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
> 

--

-- 

Vinny Abello
Network Engineer
vinny <at> tellurian.com
PGP Key Fingerprint: 3BC5 9A48 FC78 03D3 82E0  E935 5325 FBCB 0100 977A

Tellurian Networks - Global Hosting Solutions since 1995
http://www.tellurian.com | 888-835-5874 (US) | 973-300-9211 (International)

"There is no objective reality. Only that which is measured exists.
We construct reality, and only in the moment of measurement or observation." -- Niels Bohr

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users

Tobias Oetiker | 1 Feb 20:32 2008
Picon

Re: how low can it go?

Hi Aaron,

check this out

http://oss.oetiker.ch/smokeping-demo/?target=Customers.OP.james

note though, that I am not sure how accurate these mesurements are.
You may have to make sure your linux kernel is compiled
appropriately.

cheers
tobi

Today Aaron Lewis wrote:

> Before I dig into setting up smokeping, I would like to ask if I can
> monitor very small latency numbers--sub millisecond?  All nodes are in
> a single building LAN.  I need to monitor latency over time to track a
> baseline, compare changes, etc.  Everything I've looked at on Windows
> (What's Up Gold, IPCheck) won't measure smaller than 1 ms.  Is this
> because the built in PING in windows doesn't resolve lower than 1 ms?
> I notice in any Linux distro, PING gives me numbers like 0.255 ms
> which is great.  So will smokeping work at this level?
>
> If so, what is the easiest way for someone to get smokeping running?
> I can setup a dedicated machine, maybe a few years old (P4, 1GB RAM,
> 120+ hdd) polling about 100 nodes on the LAN. I would probably want to
> monitor our ISP's DNS, web, and mail servers, which would be measured
> on the more "normal" ms level. Do some distros have smokeping
> installed already?  I'm fairly new to Linux, but work as a Windows
> system admin, Net+ certified.  I'm familiar with basic CLI
> administering though SSH.  Is there something like Remote Desktop for
> Linux?  Thanks!
>
> _______________________________________________
> smokeping-users mailing list
> smokeping-users <at> lists.oetiker.ch
> https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
>
>

--

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten
http://it.oetiker.ch tobi <at> oetiker.ch ++41 62 213 9902

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users

Alexis Letessier | 1 Feb 21:21 2008
Picon

Traceroute probe

Hi everyone,

Has anyone ever tried to develop a traceroute probe that would display time series for each hop?
It would permit to detect and localize link congestion or packet loss on the path.
The main difficulty i see is about dealing with changing network paths.

For example, each periodic test would display 5 values for each hop and the graph may look like 6 different lines. Therefore each test would display 30 values.

# traceroute -f 4 -n -q 5 www.google.com
traceroute: Warning: www.l.google.com has multiple addresses; using 209.85.135.104
traceroute to www.l.google.com (209.85.135.104), 30 hops max, 40 byte packets
 4  64.233.175.234  1.894 ms  1.827 ms  1.620 ms  1.717 ms  1.720 ms
 5  72.14.232.104  18.466 ms 72.14.233.105  29.250 ms  29.406 ms 72.14.232.104  18.667 ms  18.511 ms
 6  72.14.238.128  25.928 ms  25.905 ms 72.14.233.106  25.944 ms  26.012 ms 209.85.248.248  27.828 ms
 7  72.14.239.46  28.041 ms 72.14.239.48  28.442 ms  28.388 ms 72.14.239.46  28.019 ms  27.952 ms
 8  209.85.253.26  29.345 ms 72.14.239.54  41.336 ms 209.85.253.26  29.372 ms 72.14.239.54  29.306 ms 209.85.253.26  39.806 ms
 9  209.85.135.104  28.256 ms  28.417 ms  28.313 ms  28.316 ms  28.177 ms

cheers,
Alexis

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
Fred Skrotzki | 1 Feb 22:31 2008

Re: Traceroute probe

take a look at how mtr does this

From: smokeping-users-bounces <at> lists.oetiker.ch [mailto:smokeping-users-bounces <at> lists.oetiker.ch] On Behalf Of Alexis Letessier
Sent: Friday, February 01, 2008 3:21 PM
To: smokeping-users <at> lists.oetiker.ch
Subject: [smokeping-users] Traceroute probe

Hi everyone,

Has anyone ever tried to develop a traceroute probe that would display time series for each hop?
It would permit to detect and localize link congestion or packet loss on the path.
The main difficulty i see is about dealing with changing network paths.

For example, each periodic test would display 5 values for each hop and the graph may look like 6 different lines. Therefore each test would display 30 values.

# traceroute -f 4 -n -q 5 www.google.com
traceroute: Warning: www.l.google.com has multiple addresses; using 209.85.135.104
traceroute to www.l.google.com (209.85.135.104), 30 hops max, 40 byte packets
 4  64.233.175.234  1.894 ms  1.827 ms  1.620 ms  1.717 ms  1.720 ms
 5  72.14.232.104  18.466 ms 72.14.233.105  29.250 ms  29.406 ms 72.14.232.104  18.667 ms  18.511 ms
 6  72.14.238.128  25.928 ms  25.905 ms 72.14.233.106  25.944 ms  26.012 ms 209.85.248.248  27.828 ms
 7  72.14.239.46  28.041 ms 72.14.239.48  28.442 ms  28.388 ms 72.14.239.46  28.019 ms  27.952 ms
 8  209.85.253.26  29.345 ms 72.14.239.54  41.336 ms 209.85.253.26  29.372 ms 72.14.239.54  29.306 ms 209.85.253.26  39.806 ms
 9  209.85.135.104  28.256 ms  28.417 ms  28.313 ms  28.316 ms  28.177 ms

cheers,
Alexis

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
Tobias Oetiker | 2 Feb 01:33 2008
Picon

Re: Traceroute probe

Hi Alexis,

I have an offer out to a customer who would like to see a ajax
frontend to mtr integrated with smokeping ...

don't have to contract yet, but it may get you the ability to click
on a link that looks odd and get a traceroute from the server ...
there would be no history, but it may make trouble shooting simpler
...

cheers
tobi

> take a look at how mtr does this
>
> ________________________________
>
> From: smokeping-users-bounces <at> lists.oetiker.ch
> [mailto:smokeping-users-bounces <at> lists.oetiker.ch] On Behalf Of Alexis
> Letessier
> Sent: Friday, February 01, 2008 3:21 PM
> To: smokeping-users <at> lists.oetiker.ch
> Subject: [smokeping-users] Traceroute probe
>
>
> Hi everyone,
>
> Has anyone ever tried to develop a traceroute probe that would display
> time series for each hop?
> It would permit to detect and localize link congestion or packet loss on
> the path.
> The main difficulty i see is about dealing with changing network paths.
>
> For example, each periodic test would display 5 values for each hop and
> the graph may look like 6 different lines. Therefore each test would
> display 30 values.
>
> # traceroute -f 4 -n -q 5 www.google.com
> traceroute: Warning: www.l.google.com has multiple addresses; using
> 209.85.135.104
> traceroute to www.l.google.com (209.85.135.104), 30 hops max, 40 byte
> packets
>  4  64.233.175.234  1.894 ms  1.827 ms  1.620 ms  1.717 ms  1.720 ms
>  5  72.14.232.104  18.466 ms 72.14.233.105  29.250 ms  29.406 ms
> 72.14.232.104  18.667 ms  18.511 ms
>  6  72.14.238.128  25.928 ms  25.905 ms 72.14.233.106  25.944 ms  26.012
> ms 209.85.248.248  27.828 ms
>  7  72.14.239.46  28.041 ms 72.14.239.48  28.442 ms  28.388 ms
> 72.14.239.46  28.019 ms  27.952 ms
>  8  209.85.253.26  29.345 ms 72.14.239.54  41.336 ms 209.85.253.26
> 29.372 ms 72.14.239.54  29.306 ms 209.85.253.26  39.806 ms
>  9  209.85.135.104  28.256 ms  28.417 ms  28.313 ms  28.316 ms  28.177
> ms
>
> cheers,
> Alexis
>
>
>

--

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten
http://it.oetiker.ch tobi <at> oetiker.ch ++41 62 213 9902

_______________________________________________
smokeping-users mailing list
smokeping-users <at> lists.oetiker.ch
https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users


Gmane