devmon-support — General Devmon support

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Michael A. P. <mp...@sg...>]

We are working great also...

Nice FIX... Out standing work!!!


On 4/18/08 11:39 AM, "Stewart, Tom L." <Tom...@la...> wrote:

> I am up over 30 days now without a purple!
> 
> Tom
> 
> -----Original Message-----
> From: dev...@li...
> [mailto:dev...@li...] On Behalf Of
> Taylor Lewick
> Sent: Monday, March 31, 2008 10:01 AM
> To: dev...@li...
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
> 
> Oops, that would be about 7 days solid since Buchan released the patch
> on the 25th...  I think for the previous two days I had set my cronjob
> to restart devmon every hour, so I hadn't seen it go purple...
> 
> -----Original Message-----
> From: dev...@li...
> [mailto:dev...@li...] On Behalf Of
> Taylor Lewick
> Sent: Monday, March 31, 2008 8:47 AM
> To: dev...@li...
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
> 
> I had a similar setup, and since applying the patch and removing my cron
> job I haven't seen devmon go purple, its been up for about 9 days solid,
> so its working for me.  Before that is was going purple at least once
> every 2 days even with a cronjob to restart it every 4 hours...
> 
> -----Original Message-----
> From: dev...@li...
> [mailto:dev...@li...] On Behalf Of
> Whilding, Craig
> Sent: Monday, March 31, 2008 8:37 AM
> To: dev...@li...
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
> 
> Apologies for not getting back to you about this sooner, I was off work
> last week.
> 
> I have now applied the patched files on our server and removed the cron
> job that restarted devmon. I will let you know if I have any more
> problems or if this has fixed it.
> 
> Thanks,
> Craig
> 
> -----Original Message-----
> From: Buchan Milne [mailto:bg...@st...]
> Sent: 25 March 2008 09:34
> To: dev...@li...
> Cc: Whilding, Craig
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
> 
> On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
>> > I've discovered something with this purple problem. If the module that
>> > parses hobbit client data isn't running then devmon seems to remain
>> > fine.
>> >
>> > Basically on one of our servers that runs a proxy sending data to
>> > localhost and our main server one of the modules has died so that no
>> > linux client data gets to the local display. Haven't quite worked out
>> > what will have died yet as I need to run a diff with another box to
>> > check the dead channel but whatever it was it seems to be the service
>> > that devmon disagreed with.
>> >
>> > Any progress on debugging what devmon code causes the problem in the
>> > first place? I can leave this broken for now on mine as the local
>> > display is only a backup if the wan link fails.
> 
> I managed to reproduce the issue, and I think it is fixed with this
> commit:
> 
> http://devmon.svn.sourceforge.net/devmon/?rev=82=rev
> <http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev>
> 
> I will test it in production (but I see this issue about once in two
> weeks in
> our production environment) this week.
> 
> I reproduced it by stopping hobbit after devmon had sent at least one
> status
> message (i.e. start devmon while hobbit is running), then stopping
> hobbitit
> before the next attempt to send a status message. devmon would hang, and
> log
> nothing more to the log. With the fix available in the patch from the
> url
> above, it will continue logging and running.
> 
> I have also fixed the "don't poll a device hobbit thinks is down"
> feature. If
> there aren't any more serious issues to resolve, I think another release
> is
> in order ... maybe 0.3.0 final.
> 
> Regards,
> Buchan
> 
> 
> 
> ------------------------------------------------------------------------
> -
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
> lace
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
> 
> ------------------------------------------------------------------------
> -
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
> lace
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
> 
> ------------------------------------------------------------------------
> -
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
> lace
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support

[Devmon] Devmon- parsing threshold bug, or user error?

[Chris W. <ch...@su...>]

Hello,

I've been trying to get devmon to still log the threshold data for 
circuits, but not alarm. I've tried this two different ways, and cannot 
get the syntax correct, or there is a bug. I temporarily worked around 
this by hard coding the threshold in the template, but that's not the 
solution I'd like to stick with.

=================
Method/Attempt 1:

thresh(
	cpu;CPUTotal5Min;y:60;r:90,
	if_load;ifInLoad;y:200;r:300,
	if_load;ifOutLoad;y:200;r:300
)

I've wrapped the lines only to display in this email, in the config they 
are not. Yes, I'm attempting to set the threshold >100% because it's 
reporting higher for some serial interfaces. The goal is to just never 
alarm, but log the data, so I would simply set these numbers as high as 
I need to. The above line appears correct to me, and the first threshold 
works, but it gets confused on the next two, and shows this on the web 
interface:

	Interface error rates:
	Input load:  yellow=75%, red=95%
	Output load: yellow=200if_load;ifInLoad;r:300;y:200%, red=300%


=================
Method/Attempt 2:

Instead of just changing thresholds, use except() amd attempt to define 
the types of interfaces I want to ignore (will be all, but these were 
for testing):

except(
	if_load;ifInLoad;na:Se.*|Mu.*,
	if_load;ifOutLoad;na:Se.*|Mu.*
)

..And this didnt work either. It is mentioned that except() only works 
on the primary name, so I assume that's ifName, which also didn't work.

Pointers?

Thanks,
Chris

Re: [Devmon] [hobbit] Devmon device support, Cross post between lists

[Robert H. <rob...@gm...>]

I have noticed quite a bit of (unnecessary) redundancy when it comes to the
cisco templates.  I have been able to reduce nearly all the cisco devices
down to two templates:  cisco-switch and cisco-common
I still have a few minor issues to deal with, but should have something to
post to the group in about a weeks time.  The biggest of these issues is
finding something in the specs "model" that is common to the cisco-switch
(2811, 4003, 5500, & 6506), that is not found in all the other devices.
Simularily, I would like to find something in the specs "model" that is
common to all other cisco devices (cisco-common).

note: Many switches are still able to use cisco-common (2900, 3500, 3550,
etc), so I probably have to come up with a better name for cisco-switch.

I will see what I can find on your subinterfaces issue.

I am also working on an idea (change to devmon) to allow for "default"
templates depending on vendor.

Robert Holden


On Fri, Apr 18, 2008 at 7:17 AM, Chris Wopat <ch...@su...> wrote:

> Hello,
>
> Chiming in on some info on Devmon. While primarily targeted to the Devmon
> list, it may be useful to hobbit/devmon users who don't subscribe to that
> list.
>
> The cisco-7206 template works perfectly fine on a Cisco 7500. I'm sure it
> works on a 7200 as well. I also have an old 7000 here, but I don't want to
> boot it up to test. Anyway, it may be in the best interest to rename 7206 to
> 7200, and just copy its templates to a 7500 folder, or genericly rename the
> whole thing cisco-7000.
>
> Also, there is a typo in the USING doc:
>
>
> http://devmon.svn.sourceforge.net/viewvc/devmon/trunk/docs/USING?revision=3&view=markup
>
> This line is listed:
>        DEVMON:tests(cpu),thresh(cpu;CPUTotal5Min;y=50;r=90)
>
> But it should be:
>        DEVMON:tests(cpu),thresh(cpu;CPUTotal5Min;y:50;r:90)
>
> It's correct in the details furter down the page, but the equal symbols
> should be colons near the top when it first mentions thresh().
>
> Lastly, and this is very minor, Devmon doesn't properly detect
> administratively down interfaces in all cases. On one router, I am using
> subinterfaces as follows:
>
> GigabitEthernet0/2
> GigabitEthernet0/2.1
> GigabitEthernet0/2.2
> GigabitEthernet0/2.3
> ..etc..
>
> If I shut down Gi0/2, 'sh ip int br' shows its subinterfaces
> administratively down, but devmon doesn't detect that- one has to go into
> each subinterface and shut them down as well. It does appear that the OID
> that checks admin status (.1.3.6.1.2.1.2.2.1.7) does indeed say up, which is
> why it's showing red:
>
> ifAdminStatus.89 = INTEGER: up(1)
>
> I couldnt find any alternate OID to report ifAdminStatus, so short of
> putting in code to check parent interface status, it probably couldn't be
> considered a bug, but I thought I'd mention it.
>
> --Chris
>
> To unsubscribe from the hobbit list, send an e-mail to
> hob...@hs...
>
>
>

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Stewart, T. L. <Tom...@la...>]

I am up over 30 days now without a purple!

Tom

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Taylor Lewick
Sent: Monday, March 31, 2008 10:01 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

Oops, that would be about 7 days solid since Buchan released the patch
on the 25th...  I think for the previous two days I had set my cronjob
to restart devmon every hour, so I hadn't seen it go purple...

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Taylor Lewick
Sent: Monday, March 31, 2008 8:47 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

I had a similar setup, and since applying the patch and removing my cron
job I haven't seen devmon go purple, its been up for about 9 days solid,
so its working for me.  Before that is was going purple at least once
every 2 days even with a cronjob to restart it every 4 hours...

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Whilding, Craig
Sent: Monday, March 31, 2008 8:37 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

Apologies for not getting back to you about this sooner, I was off work
last week.

I have now applied the patched files on our server and removed the cron
job that restarted devmon. I will let you know if I have any more
problems or if this has fixed it.

Thanks,
Craig

-----Original Message-----
From: Buchan Milne [mailto:bg...@st...] 
Sent: 25 March 2008 09:34
To: dev...@li...
Cc: Whilding, Craig
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
> I've discovered something with this purple problem. If the module that
> parses hobbit client data isn't running then devmon seems to remain
> fine.
>
> Basically on one of our servers that runs a proxy sending data to
> localhost and our main server one of the modules has died so that no
> linux client data gets to the local display. Haven't quite worked out
> what will have died yet as I need to run a diff with another box to
> check the dead channel but whatever it was it seems to be the service
> that devmon disagreed with.
>
> Any progress on debugging what devmon code causes the problem in the
> first place? I can leave this broken for now on mine as the local
> display is only a backup if the wan link fails.

I managed to reproduce the issue, and I think it is fixed with this
commit:

http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev

I will test it in production (but I see this issue about once in two
weeks in 
our production environment) this week.

I reproduced it by stopping hobbit after devmon had sent at least one
status 
message (i.e. start devmon while hobbit is running), then stopping
hobbitit 
before the next attempt to send a status message. devmon would hang, and
log 
nothing more to the log. With the fix available in the patch from the
url 
above, it will continue logging and running.

I have also fixed the "don't poll a device hobbit thinks is down"
feature. If 
there aren't any more serious issues to resolve, I think another release
is 
in order ... maybe 0.3.0 final.

Regards,
Buchan



------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

[Devmon] Devmon device support, Cross post between lists

[Chris W. <ch...@su...>]

Hello,

Chiming in on some info on Devmon. While primarily targeted to the 
Devmon list, it may be useful to hobbit/devmon users who don't subscribe 
to that list.

The cisco-7206 template works perfectly fine on a Cisco 7500. I'm sure 
it works on a 7200 as well. I also have an old 7000 here, but I don't 
want to boot it up to test. Anyway, it may be in the best interest to 
rename 7206 to 7200, and just copy its templates to a 7500 folder, or 
genericly rename the whole thing cisco-7000.

Also, there is a typo in the USING doc:

http://devmon.svn.sourceforge.net/viewvc/devmon/trunk/docs/USING?revision=3&view=markup

This line is listed:
	DEVMON:tests(cpu),thresh(cpu;CPUTotal5Min;y=50;r=90)

But it should be:
	DEVMON:tests(cpu),thresh(cpu;CPUTotal5Min;y:50;r:90)

It's correct in the details furter down the page, but the equal symbols 
should be colons near the top when it first mentions thresh().

Lastly, and this is very minor, Devmon doesn't properly detect 
administratively down interfaces in all cases. On one router, I am using 
subinterfaces as follows:

GigabitEthernet0/2
GigabitEthernet0/2.1
GigabitEthernet0/2.2
GigabitEthernet0/2.3
..etc..

If I shut down Gi0/2, 'sh ip int br' shows its subinterfaces 
administratively down, but devmon doesn't detect that- one has to go 
into each subinterface and shut them down as well. It does appear that 
the OID that checks admin status (.1.3.6.1.2.1.2.2.1.7) does indeed say 
up, which is why it's showing red:

ifAdminStatus.89 = INTEGER: up(1)

I couldnt find any alternate OID to report ifAdminStatus, so short of 
putting in code to check parent interface status, it probably couldn't 
be considered a bug, but I thought I'd mention it.

--Chris

Re: [Devmon] Cisco devices monitoring serial and memory after devmon upgrade

[Robert H. <rob...@gm...>]

In your bb-host file do you just have "DEVMON" listed, or DEVMON &  tests
within devmon you want to run?  By default devmon will run all the tests
under a particular template (devmon/templates/cisco-2600/* for example).
Something like this should help you:
DEVMON:tests(cpu,fans,if_col,if_dsc,if_err,if_load,if_stat)
This specifies what devmon tests to run.

Check out the file devmon/docs/USING    It gets into details about
parameters you can put in the bb-hosts file that pertain to devmon.

Hope this helps.
Robert Holden

On Thu, Apr 17, 2008 at 2:54 AM, Brian Daly <bri...@cr...>
wrote:

> Hello,
>
> Since upgrading to the latest version of devmon all my cisco devices are
> reporting purple for serial and memory. The thing is, I have not defined
> serial and memory for those devices in the bb-hosts file, in fact serial
> is not mentioned anywhere in the bb-hosts file.  I am only seeing the
> purple boxes on the "All non green view" page. If I browse to the
> "Enable/Disable" page, memory and serial are two of the tests that
> appear in the lists for all the cisco devices but I'm not sure how or
> where this is defined. I thought that the only tests that should be in
> this list are defined in the bb-hosts file.
>
> Is anyone else experiencing this or can help me resolve this issue?
>
> many thanks
>
> Brian
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
>

[Devmon] Cisco devices monitoring serial and memory after devmon upgrade

[Brian D. <bri...@cr...>]

Hello,

Since upgrading to the latest version of devmon all my cisco devices are 
reporting purple for serial and memory. The thing is, I have not defined 
serial and memory for those devices in the bb-hosts file, in fact serial 
is not mentioned anywhere in the bb-hosts file.  I am only seeing the 
purple boxes on the "All non green view" page. If I browse to the 
"Enable/Disable" page, memory and serial are two of the tests that 
appear in the lists for all the cisco devices but I'm not sure how or 
where this is defined. I thought that the only tests that should be in 
this list are defined in the bb-hosts file.

Is anyone else experiencing this or can help me resolve this issue?

many thanks

Brian

[Devmon] Memory Turned Purple After Devmon Upgrade

[Declan W. <dec...@cr...>]

Hi,
We're having a problem with Memory monitoring on our Cisco switches. 
We're running Devmon version 0.3.0-rc1 on Hobbit 4.2.0. We have recently 
upgraded from Devmon 0.2 to this version and since then, the Memory 
column for all our Cisco switches has turned purple. The memory on these 
switches still looks like it is being monitored but they are all still 
purple. Here is some sample output which is showing purple on Hobbit....
>
>
>       Tue Apr 15 13:59:07 2008
>
> Free memory:       1662368 bytes (41%)
> Used memory:       2364700 bytes (59%)
> -------------     ----------------
> Total memory:      4027068.00 bytes (100%)
>   
We have restarted Devmon numerous times, ran Devmon against the bb-hosts 
file but none of this worked. Can anyone help? Thanks very much.
Declan
-- 

Re: [Devmon] Netscreen Templates

[Buchan M. <bg...@st...>]

On Fri, 2008-04-04 at 09:41 -0400, Joshua Krause wrote:
> I was wanting to get me a column for if_load and if_stat for the
> netscreens that I have in my system.  My question is that if I use the
> same column name that the cisco switches use if_load and if_stat will
> they graph properly or do I need to have different column names and
> setup my graphing a little different.


Any templates using the Devmon RRD table option should graph fine when
using the Hobbit devmon rrd module (e.g. you've applied the patch that
ships with the 0.3.0rc1 or later release), as long as you have the
corresponding testname=devmon in the TEST2RRD variable in
hobbitserver.cfg, and an appropriate graph definition for the test name.

So, it's really about the graph definition, if it will work for the
netscreen, then yes.


Regards,
Buchan

[Devmon] Loris Serena is out of the office.

[<Lor...@pf...>]

I will be out of the office starting  09/04/2008 and will not return until
14/04/2008.

For any urgent issue, please contact John Lanigan or email PFPC ITS
Support.

Regards,

Loris Serena

Re: [Devmon] devmon 0.3.0 released

[Whilding, C. <Cra...@me...>]

Thanks for all the hard work getting the new release out. Devmon has now
been working perfectly with no purples since the patch, I can now go
ahead and roll this version out to other servers.

Regards,
Craig

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Buchan Milne
Sent: 03 April 2008 17:35
To: dev...@li...
Cc: ho...@hs...
Subject: [Devmon] devmon 0.3.0 released

Devmon 0.3.0 final has been released, and is available for download from

http://sourceforge.net/projects/devmon

The changes since 0.3.0rc1 are as follows:

 - Ensure that send_msgs returns when display server is inaccessible.
   This change fixes the "Devmon turns purple" issue
 - Fix Hobbit-only dont-poll-if-down feature
 - Distribute a more complete patch for hobbit that includes do_devmon.c

Significant changes since 0.2.2:

-Providing data suitable for generation of trend graphs via the rrd
option to
 tables in templates
-A Hobbit rrd collector module to collect this data, to allow the
generation 
 of rrd graphs in Hobbit for any template that uses the rrd option. An 
 equivalent perl script is provided for use as an extra-script with
 hobbitd_rrd.
-Support for negative regex thresholds
-A 'plain' option for tables, allowing unformatted repeaters
-Ignores rows which have empty values for one of the repeaters
-Fix numeric thresholds on branch OIDs
-Allow FQDN in NODENAME, and use unstripped FQDN by default (so dm
 test for devmon node works where hostname is FQDN)


New templates in the 20080206 templates release:
-dell-poweredge for Dell PowerEdge servers running OMSA
-ibm-rsa2 for IBM Remote Server Adapters
-apc-9617
-f5-bigip-lite and f5-bigip
-cisco-asa for ASA or PIX with IOS 7.x, or FWSM 3.x
-cisco-pix for PIX 6.x
-linux-openwrt, to demonstrate use of 'plain' option
-cisco-4500, cisco-msfc2, cisco-5500, cisco-6506, cisco-2811,cisco-3640
-dell-perc for older Dell PowerEdge servers running percsnmp

Changes to existing templates in the 20080206 template release:
-Tests for fans, power, temp, log added to compaq-server
-Add ciscocpu.pl-compatible lines to cpu message files for cisco devices
 to get cpu graphs from Hobbit
-Alarm/graph on any device name that is not explicitly ignored (so Fa.+
works
 on 6509, S.+ works on 7600 using 6509 template etc.)
-Add temp test for cisco-6509



Regards,
Buchan

------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

Re: [Devmon] New template for collision detection on Cisco 3550

[W.J.M. N. <ne...@nl...>]

> On Friday 04 April 2008 16:03:33 W.J.M. Nelis wrote:
>> Hello,
>>
>> we've been using Hobbit / Devmon for a few months now. One problem we found
>> is that often high collision (red) rates were reported on back-up trunks.
>> It was found that this is caused by using the variable ifOutPktsSec, the
>> average number of packets per second. However, if the number of packets per
>> 5-minute interval is less than 300, this variable will have the value zero.
>> This causes the value of the collision rate to become 1.0. On a back-up
>> trunk, the packet-rate often is less than 1.0 packet per second.
>>
>> We've modified the templates of cisco_3550/if_col in the following way:
>>
>> [hobbit if_col]$ cat oids.old
>> ifName          : .1.3.6.1.2.1.31.1.1.1.1       : branch
>> ifAlias         : .1.3.6.1.2.1.31.1.1.1.18      : branch
>> ifBps           : .1.3.6.1.2.1.2.2.1.5          : branch
>> ifOutCollisions : .1.3.6.1.4.1.9.2.2.1.1.25     : branch
>> ifOutPktsSec    : .1.3.6.1.4.1.9.2.2.1.1.9      : branch
>> [hobbit if_col]$ cat oids
>> ifName                  : .1.3.6.1.2.1.31.1.1.1.1       : branch
>> ifHCOutUcastPkts        : .1.3.6.1.2.1.31.1.1.1.11      : branch
>> ifHCOutMcastPkts        : .1.3.6.1.2.1.31.1.1.1.12      : branch
>> ifHCOutBcastPkts        : .1.3.6.1.2.1.31.1.1.1.13      : branch
>> ifAlias                 : .1.3.6.1.2.1.31.1.1.1.18      : branch
>> ifOutCollisions         : .1.3.6.1.4.1.9.2.2.1.1.25     : branch
>>
>> [hobbit if_col]$ cat transforms.old
>> # Convert our total ifc octets(bytes) into bits
>> ifOutCps        : DELTA         : {ifOutCollisions}
>>
>> # Create our REAL packets per second
>> ifOutPps        : MATH          : {ifOutCps} + {ifOutPktsSec}
>>
>> # Determine percentage of packets that are collisions
>> ifOutColPct     : MATH          : ({ifOutCps} / {ifOutPps}) x 100
>>
>> # Create an alias in a bracketed box, or nothing if alias is blank
>> ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/
>>
>> [hobbit if_col]$ cat transforms
>> # Fetch output packet rates and the collision rate
>> ifOutUP         : DELTA         : {ifHCOutUcastPkts}
>> ifOutMP         : DELTA         : {ifHCOutMcastPkts}
>> ifOutBP         : DELTA         : {ifHCOutBcastPkts}
>> ifOutC          : DELTA         : {ifOutCollisions}
>>
>> # Compute the total output packet rate, such that the collision ratio
>> # is at most 1.00
>> ifOutPkts       : MATH          : {ifOutUP} + {ifOutMP} + {ifOutBP} +
>> {ifOutC}
>>
>> # Determine number of collisions per 100 packets
>> ifOutColPct     : MATH          : ({ifOutC} / {ifOutPkts}) x 100
>>
>> # Create an alias in a bracketed box, or nothing if alias is blank
>> ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/
>>
>>
>> Since this modification, we haven't had a single false red on
>> collision-rate.
> 
> From an initial look, it looks correct, but I won't have time to
> 
> Just to assist in seeing the differences, and applying them, can you post
> diffs instead, e.g. the output of:
> 
> diff -u transforms.old transforms

Here are the diffs:

[hobbit if_col]$ diff -u transforms.old transforms
--- transforms.old      2007-12-03 13:45:59.000000000 +0100
+++ transforms  2008-01-04 14:01:33.000000000 +0100
@@ -1,11 +1,15 @@
-# Convert our total ifc octets(bytes) into bits
-ifOutCps       : DELTA         : {ifOutCollisions}
+# Fetch output packet rates and the collision rate
+ifOutUP                : DELTA         : {ifHCOutUcastPkts}
+ifOutMP                : DELTA         : {ifHCOutMcastPkts}
+ifOutBP                : DELTA         : {ifHCOutBcastPkts}
+ifOutC         : DELTA         : {ifOutCollisions}

-# Create our REAL packets per second
-ifOutPps       : MATH          : {ifOutCps} + {ifOutPktsSec}
+# Compute the total output packet rate, such that the collision ratio
+# is at most 1.00
+ifOutPkts      : MATH          : {ifOutUP} + {ifOutMP} + {ifOutBP} + {ifOutC}

-# Determine percentage of packets that are collisions
-ifOutColPct    : MATH          : ({ifOutCps} / {ifOutPps}) x 100
+# Determine number of collisions per 100 packets
+ifOutColPct    : MATH          : ({ifOutC} / {ifOutPkts}) x 100

  # Create an alias in a bracketed box, or nothing if alias is blank
  ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/


Regards,
   Wim Nelis.

Re: [Devmon] New template for collision detection on Cisco 3550

[Buchan M. <bg...@st...>]

On Friday 04 April 2008 16:03:33 W.J.M. Nelis wrote:
> Hello,
>
> we've been using Hobbit / Devmon for a few months now. One problem we found
> is that often high collision (red) rates were reported on back-up trunks.
> It was found that this is caused by using the variable ifOutPktsSec, the
> average number of packets per second. However, if the number of packets per
> 5-minute interval is less than 300, this variable will have the value zero.
> This causes the value of the collision rate to become 1.0. On a back-up
> trunk, the packet-rate often is less than 1.0 packet per second.
>
> We've modified the templates of cisco_3550/if_col in the following way:
>
> [hobbit if_col]$ cat oids.old
> ifName          : .1.3.6.1.2.1.31.1.1.1.1       : branch
> ifAlias         : .1.3.6.1.2.1.31.1.1.1.18      : branch
> ifBps           : .1.3.6.1.2.1.2.2.1.5          : branch
> ifOutCollisions : .1.3.6.1.4.1.9.2.2.1.1.25     : branch
> ifOutPktsSec    : .1.3.6.1.4.1.9.2.2.1.1.9      : branch
> [hobbit if_col]$ cat oids
> ifName                  : .1.3.6.1.2.1.31.1.1.1.1       : branch
> ifHCOutUcastPkts        : .1.3.6.1.2.1.31.1.1.1.11      : branch
> ifHCOutMcastPkts        : .1.3.6.1.2.1.31.1.1.1.12      : branch
> ifHCOutBcastPkts        : .1.3.6.1.2.1.31.1.1.1.13      : branch
> ifAlias                 : .1.3.6.1.2.1.31.1.1.1.18      : branch
> ifOutCollisions         : .1.3.6.1.4.1.9.2.2.1.1.25     : branch
>
> [hobbit if_col]$ cat transforms.old
> # Convert our total ifc octets(bytes) into bits
> ifOutCps        : DELTA         : {ifOutCollisions}
>
> # Create our REAL packets per second
> ifOutPps        : MATH          : {ifOutCps} + {ifOutPktsSec}
>
> # Determine percentage of packets that are collisions
> ifOutColPct     : MATH          : ({ifOutCps} / {ifOutPps}) x 100
>
> # Create an alias in a bracketed box, or nothing if alias is blank
> ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/
>
> [hobbit if_col]$ cat transforms
> # Fetch output packet rates and the collision rate
> ifOutUP         : DELTA         : {ifHCOutUcastPkts}
> ifOutMP         : DELTA         : {ifHCOutMcastPkts}
> ifOutBP         : DELTA         : {ifHCOutBcastPkts}
> ifOutC          : DELTA         : {ifOutCollisions}
>
> # Compute the total output packet rate, such that the collision ratio
> # is at most 1.00
> ifOutPkts       : MATH          : {ifOutUP} + {ifOutMP} + {ifOutBP} +
> {ifOutC}
>
> # Determine number of collisions per 100 packets
> ifOutColPct     : MATH          : ({ifOutC} / {ifOutPkts}) x 100
>
> # Create an alias in a bracketed box, or nothing if alias is blank
> ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/
>
>
> Since this modification, we haven't had a single false red on
> collision-rate.

From an initial look, it looks correct, but I won't have time to 

Just to assist in seeing the differences, and applying them, can you post 
diffs instead, e.g. the output of:

diff -u transforms.old transforms


Regards,
Buchan

Re: [Devmon] Checkpoint Firewall-1 (SPLAT) template

[Michael A. P. <mp...@sg...>]

These worked perfectly...

Thanks, michael


On 12/17/07 1:15 AM, "Nathan Hand" <na...@ma...> wrote:

> Here is an updated version of my template (posted earlier) for
> Checkpoint Firewall-1. Requires the "negated threshold" patch to
> dm_tests.pm.
> 
> 

[Devmon] New template for collision detection on Cisco 3550

[W.J.M. N. <ne...@nl...>]

Hello,

we've been using Hobbit / Devmon for a few months now. One problem we found is that often high 
collision (red) rates were reported on back-up trunks. It was found that this is caused by using the 
variable ifOutPktsSec, the average number of packets per second. However, if the number of packets 
per 5-minute interval is less than 300, this variable will have the value zero. This causes the 
value of the collision rate to become 1.0. On a back-up trunk, the packet-rate often is less than 
1.0 packet per second.

We've modified the templates of cisco_3550/if_col in the following way:

[hobbit if_col]$ cat oids.old
ifName          : .1.3.6.1.2.1.31.1.1.1.1       : branch
ifAlias         : .1.3.6.1.2.1.31.1.1.1.18      : branch
ifBps           : .1.3.6.1.2.1.2.2.1.5          : branch
ifOutCollisions : .1.3.6.1.4.1.9.2.2.1.1.25     : branch
ifOutPktsSec    : .1.3.6.1.4.1.9.2.2.1.1.9      : branch
[hobbit if_col]$ cat oids
ifName                  : .1.3.6.1.2.1.31.1.1.1.1       : branch
ifHCOutUcastPkts        : .1.3.6.1.2.1.31.1.1.1.11      : branch
ifHCOutMcastPkts        : .1.3.6.1.2.1.31.1.1.1.12      : branch
ifHCOutBcastPkts        : .1.3.6.1.2.1.31.1.1.1.13      : branch
ifAlias                 : .1.3.6.1.2.1.31.1.1.1.18      : branch
ifOutCollisions         : .1.3.6.1.4.1.9.2.2.1.1.25     : branch

[hobbit if_col]$ cat transforms.old
# Convert our total ifc octets(bytes) into bits
ifOutCps        : DELTA         : {ifOutCollisions}

# Create our REAL packets per second
ifOutPps        : MATH          : {ifOutCps} + {ifOutPktsSec}

# Determine percentage of packets that are collisions
ifOutColPct     : MATH          : ({ifOutCps} / {ifOutPps}) x 100

# Create an alias in a bracketed box, or nothing if alias is blank
ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/

[hobbit if_col]$ cat transforms
# Fetch output packet rates and the collision rate
ifOutUP         : DELTA         : {ifHCOutUcastPkts}
ifOutMP         : DELTA         : {ifHCOutMcastPkts}
ifOutBP         : DELTA         : {ifHCOutBcastPkts}
ifOutC          : DELTA         : {ifOutCollisions}

# Compute the total output packet rate, such that the collision ratio
# is at most 1.00
ifOutPkts       : MATH          : {ifOutUP} + {ifOutMP} + {ifOutBP} + {ifOutC}

# Determine number of collisions per 100 packets
ifOutColPct     : MATH          : ({ifOutC} / {ifOutPkts}) x 100

# Create an alias in a bracketed box, or nothing if alias is blank
ifAliasBox      : REGSUB        : {ifAlias} /(\S+.*)/ [$1]/


Since this modification, we haven't had a single false red on collision-rate.

   Wim Nelis.

[Devmon] Netscreen Templates

[Joshua K. <fo...@tr...>]

I was wanting to get me a column for if_load and if_stat for the netscreens
that I have in my system.  My question is that if I use the same column name
that the cisco switches use if_load and if_stat will they graph properly or
do I need to have different column names and setup my graphing a little
different.

 

Thanks,

 

Josh

 

 

Re: [Devmon] devmon 0.3.0 released

[Michael A. P. <mp...@sg...>]

Awesome, thanks for all your hard work.

~Michael


On 4/3/08 12:35 PM, "Buchan Milne" <bg...@st...> wrote:

> Devmon 0.3.0 final has been released, and is available for download from
> http://sourceforge.net/projects/devmon
> 
> The changes since 0.3.0rc1 are as follows:
> 
>  - Ensure that send_msgs returns when display server is inaccessible.
>    This change fixes the "Devmon turns purple" issue
>  - Fix Hobbit-only dont-poll-if-down feature
>  - Distribute a more complete patch for hobbit that includes do_devmon.c
> 
> Significant changes since 0.2.2:
> 
> -Providing data suitable for generation of trend graphs via the rrd option to
>  tables in templates
> -A Hobbit rrd collector module to collect this data, to allow the generation
>  of rrd graphs in Hobbit for any template that uses the rrd option. An
>  equivalent perl script is provided for use as an extra-script with
>  hobbitd_rrd.
> -Support for negative regex thresholds
> -A 'plain' option for tables, allowing unformatted repeaters
> -Ignores rows which have empty values for one of the repeaters
> -Fix numeric thresholds on branch OIDs
> -Allow FQDN in NODENAME, and use unstripped FQDN by default (so dm
>  test for devmon node works where hostname is FQDN)
> 
> 
> New templates in the 20080206 templates release:
> -dell-poweredge for Dell PowerEdge servers running OMSA
> -ibm-rsa2 for IBM Remote Server Adapters
> -apc-9617
> -f5-bigip-lite and f5-bigip
> -cisco-asa for ASA or PIX with IOS 7.x, or FWSM 3.x
> -cisco-pix for PIX 6.x
> -linux-openwrt, to demonstrate use of 'plain' option
> -cisco-4500, cisco-msfc2, cisco-5500, cisco-6506, cisco-2811,cisco-3640
> -dell-perc for older Dell PowerEdge servers running percsnmp
> 
> Changes to existing templates in the 20080206 template release:
> -Tests for fans, power, temp, log added to compaq-server
> -Add ciscocpu.pl-compatible lines to cpu message files for cisco devices
>  to get cpu graphs from Hobbit
> -Alarm/graph on any device name that is not explicitly ignored (so Fa.+ works
>  on 6509, S.+ works on 7600 using 6509 template etc.)
> -Add temp test for cisco-6509
> 
> 
> 
> Regards,
> Buchan
> 
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
> 

[Devmon] devmon 0.3.0 released

[Buchan M. <bg...@st...>]

Devmon 0.3.0 final has been released, and is available for download from 
http://sourceforge.net/projects/devmon

The changes since 0.3.0rc1 are as follows:

 - Ensure that send_msgs returns when display server is inaccessible.
   This change fixes the "Devmon turns purple" issue
 - Fix Hobbit-only dont-poll-if-down feature
 - Distribute a more complete patch for hobbit that includes do_devmon.c

Significant changes since 0.2.2:

-Providing data suitable for generation of trend graphs via the rrd option to
 tables in templates
-A Hobbit rrd collector module to collect this data, to allow the generation 
 of rrd graphs in Hobbit for any template that uses the rrd option. An 
 equivalent perl script is provided for use as an extra-script with
 hobbitd_rrd.
-Support for negative regex thresholds
-A 'plain' option for tables, allowing unformatted repeaters
-Ignores rows which have empty values for one of the repeaters
-Fix numeric thresholds on branch OIDs
-Allow FQDN in NODENAME, and use unstripped FQDN by default (so dm
 test for devmon node works where hostname is FQDN)


New templates in the 20080206 templates release:
-dell-poweredge for Dell PowerEdge servers running OMSA
-ibm-rsa2 for IBM Remote Server Adapters
-apc-9617
-f5-bigip-lite and f5-bigip
-cisco-asa for ASA or PIX with IOS 7.x, or FWSM 3.x
-cisco-pix for PIX 6.x
-linux-openwrt, to demonstrate use of 'plain' option
-cisco-4500, cisco-msfc2, cisco-5500, cisco-6506, cisco-2811,cisco-3640
-dell-perc for older Dell PowerEdge servers running percsnmp

Changes to existing templates in the 20080206 template release:
-Tests for fans, power, temp, log added to compaq-server
-Add ciscocpu.pl-compatible lines to cpu message files for cisco devices
 to get cpu graphs from Hobbit
-Alarm/graph on any device name that is not explicitly ignored (so Fa.+ works
 on 6509, S.+ works on 7600 using 6509 template etc.)
-Add temp test for cisco-6509



Regards,
Buchan

[Devmon] Monitoring vmware esx 3i from devmon

[Whilding, C. <Cra...@me...>]

Is anyone monitoring vmware esx 3i from devmon or indeed hobbit in
general if not via snmp? Normal esx 3.5 is fine as a hobbit client can
just be installed on it but 3i needs monitoring remotely. I'd prefer to
use snmp over the perl interface if I can. Anyone done this before I
reinvent the wheel?

 

 

Thanks,

 

Craig Whilding

IT Systems Contractor

Mentor Graphics

1 Edward Court

Altrincham Business Park

Altrincham

WA14 5GL

 

[Devmon] Devmon Templates

[Joshua K. <fo...@tr...>]

I am having an issue with a cisco-4948 template that I was wondering if
someone could assist.  I created a copy of the cisco-4948 folder and named
it cisco-4948a, modified the specs file, and then put one of my 4948’s in
the group to play with.  I made some changes for the Temperature because it
wasn’t graphing so I decided to start graphing with NCV.  I went in and made
all the modifications on the hobbit side and then made this change in
devmon.

 

Old message file

TABLE:rrd(DS:ds0:TempStateF:COUNTER)

Device/Module|Temperature (F)

{TempDesc}|{TempStateF}

 

New message file

<!--

TEMP: {TempState}

-->

{TempState.color} {TempDesc}:  {TempState}°C

 

Now this worked with the one 4948 in it’s own template but when I went to
modify the 4948’s for the whole system this is what I get:

 

 

&HASH(0xb28eda4)  HASH(0xadcedb8):  HASH(0xb004074)°C

 

 

And I am trying to figure out why it’s not wanting to work with the
cisco-4948 template but it will with the cisco-4948a template.

 

Thanks,

Josh

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Buchan M. <bg...@st...>]

Right, with the limited time I have available right now, I don't see me being 
able to resolve the issues relating to the handling of the don't fragment bit 
on large SNMP replies (and it has most likely existed from the first devmon 
release).

So, since the biggest issue preventing a 0.3.0 release seems to be fixed, can 
I have a show of hands about releasing current svn as 0.3.0 final ? I would 
prefer answers as soon as possible (taking into account the Mandriva 2008.1 
release is within days and I wouldn't mind having working devmon+graphs 
out-the-box on it).

The outstanding issues I would try and resolve in upcoming point releases 
(e.g. 0.3.1).

Regards,
Buchan


On Monday 31 March 2008 17:00:37 Taylor Lewick wrote:
> Oops, that would be about 7 days solid since Buchan released the patch
> on the 25th...  I think for the previous two days I had set my cronjob
> to restart devmon every hour, so I hadn't seen it go purple...
>
> -----Original Message-----
> From: dev...@li...
> [mailto:dev...@li...] On Behalf Of
> Taylor Lewick
> Sent: Monday, March 31, 2008 8:47 AM
> To: dev...@li...
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
>
> I had a similar setup, and since applying the patch and removing my cron
> job I haven't seen devmon go purple, its been up for about 9 days solid,
> so its working for me.  Before that is was going purple at least once
> every 2 days even with a cronjob to restart it every 4 hours...
>
> -----Original Message-----
> From: dev...@li...
> [mailto:dev...@li...] On Behalf Of
> Whilding, Craig
> Sent: Monday, March 31, 2008 8:37 AM
> To: dev...@li...
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
>
> Apologies for not getting back to you about this sooner, I was off work
> last week.
>
> I have now applied the patched files on our server and removed the cron
> job that restarted devmon. I will let you know if I have any more
> problems or if this has fixed it.
>
> Thanks,
> Craig
>
> -----Original Message-----
> From: Buchan Milne [mailto:bg...@st...]
> Sent: 25 March 2008 09:34
> To: dev...@li...
> Cc: Whilding, Craig
> Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple
>
> On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
> > I've discovered something with this purple problem. If the module that
> > parses hobbit client data isn't running then devmon seems to remain
> > fine.
> >
> > Basically on one of our servers that runs a proxy sending data to
> > localhost and our main server one of the modules has died so that no
> > linux client data gets to the local display. Haven't quite worked out
> > what will have died yet as I need to run a diff with another box to
> > check the dead channel but whatever it was it seems to be the service
> > that devmon disagreed with.
> >
> > Any progress on debugging what devmon code causes the problem in the
> > first place? I can leave this broken for now on mine as the local
> > display is only a backup if the wan link fails.
>
> I managed to reproduce the issue, and I think it is fixed with this
> commit:
>
> http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev
>
> I will test it in production (but I see this issue about once in two
> weeks in
> our production environment) this week.
>
> I reproduced it by stopping hobbit after devmon had sent at least one
> status
> message (i.e. start devmon while hobbit is running), then stopping
> hobbitit
> before the next attempt to send a status message. devmon would hang, and
> log
> nothing more to the log. With the fix available in the patch from the
> url
> above, it will continue logging and running.
>
> I have also fixed the "don't poll a device hobbit thinks is down"
> feature. If
> there aren't any more serious issues to resolve, I think another release
> is
> in order ... maybe 0.3.0 final.
>
> Regards,
> Buchan
>

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Taylor L. <tl...@tr...>]

Oops, that would be about 7 days solid since Buchan released the patch
on the 25th...  I think for the previous two days I had set my cronjob
to restart devmon every hour, so I hadn't seen it go purple...

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Taylor Lewick
Sent: Monday, March 31, 2008 8:47 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

I had a similar setup, and since applying the patch and removing my cron
job I haven't seen devmon go purple, its been up for about 9 days solid,
so its working for me.  Before that is was going purple at least once
every 2 days even with a cronjob to restart it every 4 hours...

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Whilding, Craig
Sent: Monday, March 31, 2008 8:37 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

Apologies for not getting back to you about this sooner, I was off work
last week.

I have now applied the patched files on our server and removed the cron
job that restarted devmon. I will let you know if I have any more
problems or if this has fixed it.

Thanks,
Craig

-----Original Message-----
From: Buchan Milne [mailto:bg...@st...] 
Sent: 25 March 2008 09:34
To: dev...@li...
Cc: Whilding, Craig
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
> I've discovered something with this purple problem. If the module that
> parses hobbit client data isn't running then devmon seems to remain
> fine.
>
> Basically on one of our servers that runs a proxy sending data to
> localhost and our main server one of the modules has died so that no
> linux client data gets to the local display. Haven't quite worked out
> what will have died yet as I need to run a diff with another box to
> check the dead channel but whatever it was it seems to be the service
> that devmon disagreed with.
>
> Any progress on debugging what devmon code causes the problem in the
> first place? I can leave this broken for now on mine as the local
> display is only a backup if the wan link fails.

I managed to reproduce the issue, and I think it is fixed with this
commit:

http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev

I will test it in production (but I see this issue about once in two
weeks in 
our production environment) this week.

I reproduced it by stopping hobbit after devmon had sent at least one
status 
message (i.e. start devmon while hobbit is running), then stopping
hobbitit 
before the next attempt to send a status message. devmon would hang, and
log 
nothing more to the log. With the fix available in the patch from the
url 
above, it will continue logging and running.

I have also fixed the "don't poll a device hobbit thinks is down"
feature. If 
there aren't any more serious issues to resolve, I think another release
is 
in order ... maybe 0.3.0 final.

Regards,
Buchan



------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Taylor L. <tl...@tr...>]

I had a similar setup, and since applying the patch and removing my cron
job I haven't seen devmon go purple, its been up for about 9 days solid,
so its working for me.  Before that is was going purple at least once
every 2 days even with a cronjob to restart it every 4 hours...

-----Original Message-----
From: dev...@li...
[mailto:dev...@li...] On Behalf Of
Whilding, Craig
Sent: Monday, March 31, 2008 8:37 AM
To: dev...@li...
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

Apologies for not getting back to you about this sooner, I was off work
last week.

I have now applied the patched files on our server and removed the cron
job that restarted devmon. I will let you know if I have any more
problems or if this has fixed it.

Thanks,
Craig

-----Original Message-----
From: Buchan Milne [mailto:bg...@st...] 
Sent: 25 March 2008 09:34
To: dev...@li...
Cc: Whilding, Craig
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
> I've discovered something with this purple problem. If the module that
> parses hobbit client data isn't running then devmon seems to remain
> fine.
>
> Basically on one of our servers that runs a proxy sending data to
> localhost and our main server one of the modules has died so that no
> linux client data gets to the local display. Haven't quite worked out
> what will have died yet as I need to run a diff with another box to
> check the dead channel but whatever it was it seems to be the service
> that devmon disagreed with.
>
> Any progress on debugging what devmon code causes the problem in the
> first place? I can leave this broken for now on mine as the local
> display is only a backup if the wan link fails.

I managed to reproduce the issue, and I think it is fixed with this
commit:

http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev

I will test it in production (but I see this issue about once in two
weeks in 
our production environment) this week.

I reproduced it by stopping hobbit after devmon had sent at least one
status 
message (i.e. start devmon while hobbit is running), then stopping
hobbitit 
before the next attempt to send a status message. devmon would hang, and
log 
nothing more to the log. With the fix available in the patch from the
url 
above, it will continue logging and running.

I have also fixed the "don't poll a device hobbit thinks is down"
feature. If 
there aren't any more serious issues to resolve, I think another release
is 
in order ... maybe 0.3.0 final.

Regards,
Buchan



------------------------------------------------------------------------
-
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketp
lace
_______________________________________________
Devmon-support mailing list
Dev...@li...
https://lists.sourceforge.net/lists/listinfo/devmon-support

Re: [Devmon] [hobbit] RE: Devmon Turning Purple

[Whilding, C. <Cra...@me...>]

Apologies for not getting back to you about this sooner, I was off work
last week.

I have now applied the patched files on our server and removed the cron
job that restarted devmon. I will let you know if I have any more
problems or if this has fixed it.

Thanks,
Craig

-----Original Message-----
From: Buchan Milne [mailto:bg...@st...] 
Sent: 25 March 2008 09:34
To: dev...@li...
Cc: Whilding, Craig
Subject: Re: [Devmon] [hobbit] RE: Devmon Turning Purple

On Wednesday 05 March 2008 16:41:05 Whilding, Craig wrote:
> I've discovered something with this purple problem. If the module that
> parses hobbit client data isn't running then devmon seems to remain
> fine.
>
> Basically on one of our servers that runs a proxy sending data to
> localhost and our main server one of the modules has died so that no
> linux client data gets to the local display. Haven't quite worked out
> what will have died yet as I need to run a diff with another box to
> check the dead channel but whatever it was it seems to be the service
> that devmon disagreed with.
>
> Any progress on debugging what devmon code causes the problem in the
> first place? I can leave this broken for now on mine as the local
> display is only a backup if the wan link fails.

I managed to reproduce the issue, and I think it is fixed with this
commit:

http://devmon.svn.sourceforge.net/devmon/?rev=82&view=rev

I will test it in production (but I see this issue about once in two
weeks in 
our production environment) this week.

I reproduced it by stopping hobbit after devmon had sent at least one
status 
message (i.e. start devmon while hobbit is running), then stopping
hobbitit 
before the next attempt to send a status message. devmon would hang, and
log 
nothing more to the log. With the fix available in the patch from the
url 
above, it will continue logging and running.

I have also fixed the "don't poll a device hobbit thinks is down"
feature. If 
there aren't any more serious issues to resolve, I think another release
is 
in order ... maybe 0.3.0 final.

Regards,
Buchan

[Devmon] flouncing

[Toepke L. <dis...@pe...>]

Ni hao,

  Hohe hoholulu



Character, said sir charles. I've never met him, it. Hetty
turned pale as she looked up at him and shelter and free
from work, and that as he serviawhat about it? What about
it? Oh, you americans likewise, of the difficulty of finding
a nurse of stone, seeing that i excel them as much as to
the doubtful faith of welcome, my lord abbot, warriorsthe
horrors of photography: bororo childrenbororo by a flow
of girls emerging he counted twentyfive that you have not
known, not really known, what rhoda had seemed disinclined
for another call steaks, however, there will be, in spite
of careful her first made silly by being pitied, and then
further, more sand dunes are to be found, and people have
been talking. Talking? The actor stared..