A problem occurred on one of our IP internet submaps. The ethernet LAN
was disconnected from our Netview computer. Normally all interface
symbols for that LAN go red but on this occasion most stayed green and
the parent node symbols also stayed green for that segment.
I did a demand poll on one node and got the following output:
Interface 131.0.0.1 (currently up)
Interface 133.0.0.1 (currently up)
131.0.0.1 (fta) is UP Responded to ping
133.0.0.1 (tu) is DOWN Timed Out
I expected the 133.0.0.1 interface to go red after this but it still
stayed green. As you can see the node is attached to two LANs, 131 and
133. The demand poll was repeated a number of times but the colour did
not change.
I was wondering what the output above corresponds to in the inner
working of netmon. As I understand it, the demand poll command sends a
trap to netmon which carries out the poll.
What's happening when the (currently up) text is output, is this just
giving the current state of the interface in the toplogy database before
doing a ping?
The second pair of lines look like the result of a ping status poll
rather than a SNMP status poll. We have not set up netmon.seed to use a
SNMP status poll for any of our nodes.
The polling was configured to poll every 3 seconds with 2 retries. Does
the ping status poll use this retry setting or is only one ping sent?
We got around this problem by clearing the database and allowing it to
re-populate itself. The problem has not occurred since. We got around to
reading the Diagnosis Guide and found how to switch on netmon trace. I
wish we knew this when the problem was occurring!
All our equipment is Compaq DS10 running TRU 64 UNIX.
I am trying to work out what might hae caused our problem as well as
understand how netmon works in this respect so thanks for any replies.
--
Robin
email: robin.james@thalesatm.com
tel: +44 (0) 1633-862020
fax: +44 (0) 1633-868313
|