Re: [nv-l] Routers in Isolated Subnet
2004-12-13 22:11:11
If I understand the question correctly,
then in most cases James' description is accurate. For the routers between
Netview and the subnet, you will get a Router Marginal event, which is
meant to indicate that they have at least one interface up and one down,
and that down one is a problem. For the remote subnet you get subnet unreachable,
and for routers beyond that, you get Router Unreachable, meaning they are
only victims. The Router Marginal events are therefore the most important
ones. There will be Router Unreachable event, and yes, they indicate the
impact of the Router Marginal problems.
There is a case in which you get a Router
Marginal for a nearby router when it is blameless. For serial interfaces,
when there is a remote failure, the nearby router will shut its end down
in response. Netview will report that nearby router as Router Marginal,
and the serial network as Network Unreachable and the remote, failing router
as Router Unreachable. This suggests that the failure was an interface
on the near router, when the problem was really the failure of the remote
router. This is probably not the case in the arrangement you are
describing, however.
Is your concern that with RFI enabled
you won't get the Node Down event for the servers on that subnet, and the
servers are what you are really worried about? You kind of have to focus
on the root cause events, the router problems, as a trade off for all of
the event suppression you get with RFI.
If you have Switch Analyzer installed,
it can generate an impact report for a given router or layer2 device.
I hope that throwing all of these words
out there will help. If not, digest this and ask again.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
(248) 552-4968 Voicemail, Fax, Pager
James Shanks/Raleigh/IBM@IBMUS
Sent by: owner-nv-l AT lists.us.ibm DOT com
12/10/2004 10:18 AM
|
To
| nv-l AT lists.us.ibm DOT com
|
cc
|
|
Subject
| Re: [nv-l] Routers in Isolated
Subnet |
|
I was hoping that someone with more experience with RFI
would try to answer you on this, but since no one has, here's my two cents.
First, there is no list nor utility that will tell you what routers netmon
currently thinks are unreachable. That might be nice enhancement to netmon
someday, but it is not available today.
But, if you have a router (A) connected to a subnet, and you have another
router (B) connected to that same subnet, and the only path to router B
is across the subnet and out via router A, then if A has a problem, you
should get
(a) either an Interface Down event if just the connecting interface is
lost, and a Router Marginal for A, or
(b) a Router Down event for router A if the entire thing is lost.
But in either case you should get a Router Unreachable event for B, and
a Subnet Unreachable event for the subnet between them. At least that is
my understanding, though I must add that I don't work on netmon. If I have
gotten this wrong, then perhaps someone else will correct me.
Now as for you admins, I have to ask why they are not planning to use the
web client when they login from their remote location?
That's exactly what it is for, so you can use the map and all the attendant
NetView functions from a remote location. All you need a is browser.
If they aren't going to do that, then the only thing I can think of for
you to do to get a list of routers which are unreachable would be to either
(1) create a smartset for isRouter and IP Status Unreachable and then use
nvUtil to query it
(2) create a report for the same thing using nvdbformat and run that
You could then email that to the admins, but putting it in a pager notice
itself would likely make the message very large.
NetView has a map precisely so that operators can get an idea of the status
of the network visually, so they can drill down and see the source of the
problem, rather than be told about it descriptively. So I still think having
them log in remotely via the web client is a much better idea. And I know
other NetView users are doing exactly this. Some of their operators take
a laptop with them wherever they go, just so they can log in from wherever
they are when the pager goes off.
Anyone else?
James Shanks
Level 3 Support for Tivoli NetView for UNIX and Windows
Tivoli Software / IBM Software Group
SIXFC AT pjm DOT com
SIXFC AT pjm DOT com
Sent by: owner-nv-l AT lists.us.ibm DOT com
12/09/2004 12:22 PM
|
|
We have parallel paths between critical servers. If a router on one path
is down, the admins will be notified. If a router on the second path goes
down, they will also be notified. Additionally, with both paths down, the
admins want to know what routers are now isolated. Will there be a router
unreachable event for each one, following or preceding the subnet unreachable?
How can we relate them to the source problem? Does RFI keep a list of the
relationships or have a utility to determine them? Do I have access to
it?
After a pager alert, these folks will most
likely login from a remote location and open their e-mail. That's where
we're trying to put the information and why we're not telling them to look
at the map.
Thanks.
Chris Six
PJM Interconnection, llc
sixfc AT pjm DOT com
|
|
|