Xymon Mailing List Archive search

howto demotool

12 messages in this thread

list Phil Wild · Thu, 3 Jan 2008 00:23:32 +0900 ·
Hello hobbit users

Has anyone used demotool?

I think this would be a great pre production testing tool as well as useful
for training.

I can't find any documentation, can some one point me in the right direction
or give me a short howto?

when I run the tool on the command line I get

./demotool
(Re)loading config
Segmentation fault

Thanks

Phil
list Josh Luthman · Wed, 2 Jan 2008 10:33:59 -0500 ·
http://www.hobbitmon.com/hobbiton/2005/10/msg00109.html

Other:
* A new utility "demotool" can be used to simulate a number of
  servers to Hobbit. This may be useful when demonstrating
  Hobbit to new users. Note: This is not included in the
  default build - to build it, run "make demo-build".

Maybe you have to build it?  I'm shooting in the dark here =P
quoted from Phil Wild


On 1/2/08, Phil Wild <user-e365c1418192@xymon.invalid> wrote:
Hello hobbit users

Has anyone used demotool?

I think this would be a great pre production testing tool as well as
useful for training.

I can't find any documentation, can some one point me in the right
direction or give me a short howto?

when I run the tool on the command line I get

./demotool
(Re)loading config
Segmentation fault

Thanks

Phil
-- 

Josh Luthman
Office: XXX-XXX-XXXX
Direct: XXX-XXX-XXXX
XXXX Wayne St
Suite XXXX
Troy, OH XXXXX

Those who don't understand UNIX are condemned to reinvent it, poorly.
--- Henry Spencer
list Phil Wild · Thu, 3 Jan 2008 09:08:25 +0900 ·
Hi Josh/mailinglist,

Nope, I compiled demotool as per release notes. It appears to be looking for
a config file of sorts but I could not find any documentation about what it
might be looking for. I took a guess at a bb-hosts file but that did not
work.

Cheers

Phil
quoted from Josh Luthman


On 03/01/2008, Josh Luthman <user-4c45a83f15cb@xymon.invalid> wrote:
http://www.hobbitmon.com/hobbiton/2005/10/msg00109.html

Other:
* A new utility "demotool" can be used to simulate a number of

  servers to Hobbit. This may be useful when demonstrating
  Hobbit to new users. Note: This is not included in the
  default build - to build it, run "make demo-build".

Maybe you have to build it?  I'm shooting in the dark here =P


On 1/2/08, Phil Wild <user-e365c1418192@xymon.invalid> wrote:
Hello hobbit users

Has anyone used demotool?

I think this would be a great pre production testing tool as well as
useful for training.

I can't find any documentation, can some one point me in the right
direction or give me a short howto?

when I run the tool on the command line I get

./demotool
(Re)loading config
Segmentation fault

Thanks

Phil
--
Josh Luthman
Office: XXX-XXX-XXXX
Direct: XXX-XXX-XXXX
XXXX Wayne St
Suite XXXX
Troy, OH XXXXX

Those who don't understand UNIX are condemned to reinvent it, poorly.
--- Henry Spencer

-- 

Tel: XXXX XXX XXX
Fax: XXXX XXX XXX
email: user-e365c1418192@xymon.invalid
list Vernon Everett · Thu, 3 Jan 2008 10:52:07 +0900 ·
Hi all
 
First off, a great New Year to you all, and a special thank-you to
Henrik for his fantastic work thought the year.
I would also extend that thank-you to every single person who
contributed in a meaningful way to this list.
 
Now, back to business.
 
Hobbit has a beautiful facility to calculate uptime stats.
But PHBs, being what they are, want more.
Is there a way to flag an outage as "scheduled" or similar, and not
include it in the uptime stats calculation?
 
Example 
We had to reboot a system for patching, but it was scheduled with the
business, and a suitable maintenance window was agreed.
All apps were downed, and the system rebooted. 
This is not classified as an outage.
 
Later in the month there was an electrical fault on the UPS, and a
server went down.
This is a real outage and needs to be included in downtime stats.
 
How do we do this?
Do we need to disable monitoring before we begin maintenance, or is
there a way to flag the outage retrospecively?
It would be good to be able to flag the scheduled outage as "not a real
outage" after the fact, because during the outage, we use Hobbit to keep
tabs on what goes down and comes up and when.
 
Regards
     Vernon
 

NOTICE: This email and any attachments are confidential. 
They may contain legally privileged information or 
copyright material. You must not read, copy, use or 
disclose them without authorisation. If you are not an 
intended recipient, please contact us at once by return 
email and then delete both messages and all attachments.
list Henrik Størner · Thu, 3 Jan 2008 07:29:44 +0100 ·
quoted from Vernon Everett
On Thu, Jan 03, 2008 at 10:52:07AM +0900, Everett, Vernon wrote:
 
Hobbit has a beautiful facility to calculate uptime stats.
But PHBs, being what they are, want more.
Is there a way to flag an outage as "scheduled" or similar, and not
include it in the uptime stats calculation?
There is, but you have to do it in advance: By disabling the host or
status that is planned to be down.
quoted from Vernon Everett
It would be good to be able to flag the scheduled outage as "not a real
outage" after the fact, because during the outage, we use Hobbit to keep
tabs on what goes down and comes up and when.
I know, I've had the same request from my own users.


Henrik
list Henrik Størner · Thu, 3 Jan 2008 07:38:50 +0100 ·
quoted from Phil Wild
On Thu, Jan 03, 2008 at 12:23:32AM +0900, Phil Wild wrote:
Has anyone used demotool?

I think this would be a great pre production testing tool as well as useful
for training.

I can't find any documentation, can some one point me in the right direction
or give me a short howto?
It's completely undocumented, as You've found out.

The configuration is actually a directory structure, by default in
/etc/hdemo, but you can change that with the --confdir=/home/foo/democonf
option.

There's a sample configuration on
http://www.hswn.dk/~henrik/demotool-config.tar.gz


Regards,
Henrik
list Vernon Everett · Thu, 3 Jan 2008 15:52:52 +0900 ·
Hi Henrik

Thanks for the reply.
Unfortunately, it wasn't quite was I was hoping for. :-)

Can I add this to the features wish-list?

Regards
    Vernon
quoted from Henrik Størner

-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] Sent: Thursday, 3 January 2008 3:30 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Hobbit downtime stats

On Thu, Jan 03, 2008 at 10:52:07AM +0900, Everett, Vernon wrote:
 Hobbit has a beautiful facility to calculate uptime stats.
But PHBs, being what they are, want more.
Is there a way to flag an outage as "scheduled" or similar, and not include it in the uptime stats calculation?
There is, but you have to do it in advance: By disabling the host or
status that is planned to be down.
It would be good to be able to flag the scheduled outage as "not a real outage" after the fact, because during the outage, we use Hobbit to keep tabs on what goes down and comes up and when.
I know, I've had the same request from my own users.


Henrik


NOTICE: This email and any attachments are confidential. They may contain legally privileged information or copyright material. You must not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages and all attachments.
list Phil Wild · Thu, 3 Jan 2008 16:15:59 +0900 ·
Thanks Henrik,

Hobbit is such a great peice of software btw...

I have untared the directory structure and have had a look. Am I right
in assuming that the messages just get sent to the server over and
over again and that there is no design in the demo tool to simulate
faults?

Cheers

Phil
quoted from Henrik Størner

On 03/01/2008, Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:
On Thu, Jan 03, 2008 at 12:23:32AM +0900, Phil Wild wrote:
Has anyone used demotool?

I think this would be a great pre production testing tool as well as useful
for training.

I can't find any documentation, can some one point me in the right direction
or give me a short howto?
It's completely undocumented, as You've found out.

The configuration is actually a directory structure, by default in
/etc/hdemo, but you can change that with the --confdir=/home/foo/democonf
option.

There's a sample configuration on
http://www.hswn.dk/~henrik/demotool-config.tar.gz


Regards,
Henrik

-- 
Tel: XXXX XXX XXX
Fax: XXXX XXX XXX
email: user-e365c1418192@xymon.invalid
list Vernon Everett · Thu, 3 Jan 2008 16:23:04 +0900 ·
While on the subject of feature wishlists, how about this one?

We have servers with file systems on them that are completely user-land.
The OS and related stuff is all our responsibility, and we look after
it, but the apps file systems are looked after by the users.
If it fills up and the apps choke, it's not our issue. If the OS chokes, it becomes our problem.

We want OS file system disk alerts, but don't give a rat's about the
user-land stuff.
The users want alert on the user application file systems, but neither
care, nor do they need to know, about the OS file systems.

Currenly, all alerts are of the form
HOST=hostname SERVICE=service
	MAIL user-2678e9aecb95@xymon.invalid

And all disk alerts go to that person.
We can set up multiple people, but all disk alerts go to all.

How about a way of splitting that, so we have
HOST=appserver SERVICE=disk
	/apps/vol1	MAIL user-18b8c29f8e77@xymon.invalid
	/apps/vol2	MAIL user-18b8c29f8e77@xymon.invalid
	/usr		MAIL user-9fcad6e0c028@xymon.invalid 	/		MAIL user-9fcad6e0c028@xymon.invalid 
Or even 
HOST=appserver SERVICE=disk:/apps/vol1	
	MAIL user-18b8c29f8e77@xymon.invalid
HOST=appserver SERVICE=disk:/apps/vol2	
	MAIL user-18b8c29f8e77@xymon.invalid
HOST=appserver SERVICE=disk:/
	MAIL user-9fcad6e0c028@xymon.invalid
HOST=appserver SERVICE=disk:/usr
	MAIL user-9fcad6e0c028@xymon.invalid

Something like that could be useful.
quoted from Vernon Everett

Regards
    Vernon


-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] Sent: Thursday, 3 January 2008 3:30 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Hobbit downtime stats

On Thu, Jan 03, 2008 at 10:52:07AM +0900, Everett, Vernon wrote:
 Hobbit has a beautiful facility to calculate uptime stats.
But PHBs, being what they are, want more.
Is there a way to flag an outage as "scheduled" or similar, and not include it in the uptime stats calculation?
There is, but you have to do it in advance: By disabling the host or
status that is planned to be down.
It would be good to be able to flag the scheduled outage as "not a real outage" after the fact, because during the outage, we use Hobbit to keep tabs on what goes down and comes up and when.
I know, I've had the same request from my own users.


Henrik


NOTICE: This email and any attachments are confidential. They may contain legally privileged information or copyright material. You must not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages and all attachments.
list Henrik Størner · Thu, 3 Jan 2008 10:14:59 +0100 ·
quoted from Phil Wild
On Thu, Jan 03, 2008 at 04:15:59PM +0900, Phil Wild wrote:
I have untared the directory structure and have had a look. Am I right
in assuming that the messages just get sent to the server over and
over again and that there is no design in the demo tool to simulate
faults?
Correct, what you'd do is to run it and then change the files to
simulate the various kinds of failures. E.g. change the
demohost/client_df file to show a filesystem being full. Or remove
the smtp/listen file to stop responding to that service.


Henrik
list Henrik Størner · Thu, 3 Jan 2008 10:22:04 +0100 ·
quoted from Vernon Everett
On Thu, Jan 03, 2008 at 04:23:04PM +0900, Everett, Vernon wrote:
While on the subject of feature wishlists, how about this one?

We have servers with file systems on them that are completely user-land.
The OS and related stuff is all our responsibility, and we look after
it, but the apps file systems are looked after by the users.
If it fills up and the apps choke, it's not our issue. 
If the OS chokes, it becomes our problem.
[wants alerts to go to different people, depending on what filesystem is
 filling up]

You do that with the GROUP setting. In hobbit-clients.cfg, configure the
thresholds for the disks with a GROUP setting:

   HOST=appserver
      DISK /apps/vol1 80 90 GROUP=userfiles

Then in hobbit-alerts.cfg you use the GROUP value to direct alerts:

   GROUP=userfiles
   	MAIL user-18b8c29f8e77@xymon.invalid

You can of course combine the GROUP setting in hobbit-alerts.cfg with
other criteria, e.g.

   GROUP=userfiles
        MAIL user-325d01a66333@xymon.invalid HOST=appserver1
        MAIL user-e6bcb677a51f@xymon.invalid HOST=appserver2


Regards,
Henrik
list Vernon Everett · Fri, 4 Jan 2008 08:49:14 +0900 ·
 Absolutely brilliant.
Thanks

Regards
    Vernon
quoted from Henrik Størner


[wants alerts to go to different people, depending on what filesystem is
filling up]

You do that with the GROUP setting. In hobbit-clients.cfg, configure the
thresholds for the disks with a GROUP setting:

   HOST=appserver
      DISK /apps/vol1 80 90 GROUP=userfiles

Then in hobbit-alerts.cfg you use the GROUP value to direct alerts:

   GROUP=userfiles
   	MAIL user-18b8c29f8e77@xymon.invalid

You can of course combine the GROUP setting in hobbit-alerts.cfg with
other criteria, e.g.

   GROUP=userfiles
        MAIL user-325d01a66333@xymon.invalid HOST=appserver1
        MAIL user-e6bcb677a51f@xymon.invalid HOST=appserver2


Regards,
Henrik


NOTICE: This email and any attachments are confidential. They may contain legally privileged information or copyright material. You must not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages and all attachments.