Re: [Bigsister-general] Issue with generating events

["Backup e-mail" <suppressed>]

Thanks for the advice. Unfortunately it does not seem to work.
  Here is my new bb_event_generator.cfg:
  ++++++++++++++++++
  *.*     prio=40 down=yellow up=green delay=2 \
        check="$host.conn and ((not '$router') or $router.conn)" \
        norepeat=5 keep=1 suppressed repeatprio=10 \
        pager=/usr/share/bigsister/bin/page msgmax=60

  *.conn  prio=80 down=purple up=green delay=2 \
        norepeat=5 keep=1 suppressed repeatprio=10 \
        pager=/usr/share/bigsister/bin/page msgmax=60 \
        check="$host.conn and ((not '$router') or $router.conn)"

  *.cpu   prio=20 check="1" delay=0
*.disk  prio=60 check="1" delay=0
*.msgs  check="1" delay=0
++++++++++++++++++++++++

Peter Varlien <suppressed> wrote:
  On Tue, 13 Dec 2005 19:05:25 +0100, Backup e-mail
wrote:

> I have three Linux boxes somewhere on the Internet. They have SSH 
> channels with forwarding of the 1984 port to my MONITORing box (also 
> LINUX) on my office LAN.
> To simulate loss of connection it's easy, I just stop the SSH channel. 
> Which I do and after about 10 min the light goes from green to purple. I 
> wait 20, 30, ...hours but no mail and no pager script is launched. Asfar  
> as the MONITORing station is concerned, the machine from whom doesnot  
> receive is a DOWN machine, isn't that so?
> If that is so I thought the "down=yellow"directive will generate me an

purple is less severe than yellow. You need down=purple to get alarms for
a test that goes purple.
The documentation also describes a directive, %Autoconn the treats
no report for 15 minutes as meaning the the .conn test is red, and
receiving a report sets .conn to green. I haven't been able to get
this to work, however.

> event, or I haven't understood the logic. For the moment I do not have 
> Windows machine to monitor.
> Thanks for shining some light on this matter,
> Costantino
>
> Kai Schaetzl wrote:
> Backup e-mail wrote on Tue, 13 Dec 2005 05:59:28 -0800 (PST):
>
> > I can see the corresponding light changing from green to purple but 
> > that's it.
>
> The node isn't necessarily down when it's purple. Purple means "no info 
> sent for
> the last ten minutes" or so. The agent could just have died. I don't 
> think that
> purple will get an alarm message with down=yellow or down=red.
>
> Furthermore, if you are on Windows you have to do something else.
>
> Kai


Peter

--
Peter Værlien
Fritz Aabakkens vei 17, 7072 Heimdal, Norway
Telephone: 917 69 384, E-Mail: suppressed
http://varlien.home.online.no




			
---------------------------------
Yahoo! Shopping
 Find Great Deals on Holiday Gifts at Yahoo! Shopping

Thanks for the advice. Unfortunately it does not seem to work.

Here is my new bb_event_generator.cfg:

++++++++++++++++++

*.*     prio=40 down=yellow up=green delay=2 \
        check="$host.conn and ((not '$router') or $router.conn)" \
        norepeat=5 keep=1 suppressed.. repeatprio=10 \
        pager=/usr/share/bigsister/bin/page msgmax=60

 

*.conn  prio=80 down=purple up=green delay=2 \
        norepeat=5 keep=1 suppressed.. repeatprio=10 \
        pager=/usr/share/bigsister/bin/page msgmax=60 \
        check="$host.conn and ((not '$router') or $router.conn)"

 

*.cpu   prio=20 check="1" delay=0
*.disk  prio=60 check="1" delay=0
*.msgs  check="1" delay=0
++++++++++++++++++++++++

Peter Varlien <suppressed> wrote:

On Tue, 13 Dec 2005 19:05:25 +0100, Backup e-mail
wrote:

> I have three Linux boxes somewhere on the Internet. They have SSH
> channels with forwarding of the 1984 port to my MONITORing box (also
> LINUX) on my office LAN.
> To simulate loss of connection it's easy, I just stop the SSH channel.
> Which I do and after about 10 min the light goes from green to purple. I
> wait 20, 30, ...hours but no mail and no pager script is launched. As
> far as the MONITORing station is concerned, the machine from whom does
> not receive is a DOWN machine, isn't that so?
> If that is so I thought the "down=yellow"directive will generate me an

purple is less severe than yellow. You need down=purple to get alarms for
a test that goes purple.
The documentation also describes a directive, %Autoconn the treats
no report for 15 minutes as meaning the the .conn test is red, and
receiving a report sets .conn to green. I haven't been able to get
this to work, however.

> event, or I haven't understood the logic. For the moment I do not have
> Windows machine to monitor.
> Thanks for shining some light on this matter,
> Costantino
>
> Kai Schaetzl wrote:
> Backup e-mail wrote on Tue, 13 Dec 2005 05:59:28 -0800 (PST):
>
>> I can see the corresponding light changing from green to purple but
>> that's it.
>
> The node isn't necessarily down when it's purple. Purple means "no info
> sent for
> the last ten minutes" or so. The agent could just have died. I don't
> think that
> purple will get an alarm message with down=yellow or down=red.
>
> Furthermore, if you are on Windows you have to do something else.
>
> Kai
>


Peter

--
Peter Værlien
Fritz Aabakkens vei 17, 7072 Heimdal, Norway
Telephone: 917 69 384, E-Mail: suppressed
http://varlien.home.online.no

---

Yahoo! Shopping
Find Great Deals on Holiday Gifts at Yahoo! Shopping

Attachment: sanitizer.log
Description: Binary data