Hi,

I'm finding that shadowing is suppressing notification for failed servers that are 
plugged into a switch that's also being monitored by InterMapper.

If the server is powered off, or the Ethernet cable is unplugged, the switch reports 
that the link is down.  InterMapper dutifully marks the link as down and shadows the 
server, and no notification is sent.  There's no notification about the failed link, 
either, so I never get notified at all.

I'm not sure if this is a bug or not - this is how shadowing is supposed to work.  
However, the emergent behavior is clearly undesireable.  

I can't find a way to set a notification on a link or a network, which would fix the 
problem.  Maybe I missed it?

I suppose I could have InterMapper notify me each time the switch has an alarm 
condition, but there are usually a number of unimportant devices plugged into the 
switches besides the server(s) I am monitoring.  I only want notifications to happen 
on the important devices.  I suppose I could have a separate map with only the 
notification-worthy devices, but really - shouldn't I be able to do per-device 
notification?

Any hints would be welcome.

As an enhancement request, perhaps the notification algorithm could be modified to 
check the notification status of every shadowed device, and send only the minimum 
number of pages to alert everyone.

For example, let's say that there's a switch A1, managed by Ann, with three servers 
B1, B2, and C1 plugged into it, managed by Beth and Charlene.  Each of these devices 
have a pager notification attached to them to page the device's manager.

If someone turns off server B1, Beth should be paged with the message that her 
server's behind a dead link.
If someone turns off servers B1 and B2, Beth should be paged twice, once for each 
server.  (Two separate events.)
If someone turns off all three servers, Beth should receive two pages and Charlene 
should receive one.
If someone turns off the switch, then Ann, Beth, and Charlene should each get one 
page.  Ann is told that the switch is down, and Beth and Charlene are told that they 
each have at least one server behind the dead switch.  Ann gets paged first.

Note that if Ann is managing everything, only one page is sent to Ann saying that the 
switch is down.  Perhaps the page should include a list of every device that's down, 
up to the size of one pager message.  Perhaps it should just say "other devices 
affected".

Would this work for everyone else?  Too much paging?  Stupid idea?  Outside the scope 
of InterMapper?

Thanks,

Doug


-- 
Doug Weathers, Network Administrator
St. Charles Medical Center




____________________________________________________________________
List archives: http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [EMAIL PROTECTED]

Reply via email to