[evla-sw-discuss] Dropped Alerts clearm messages

[James Robnett]

  According to Matt we dropped 3 Elevation alert clear messages at
approximately 21:01 MDT (03:01 UT).   I'm assuming it hasn't happened
since.

  By the time I looked at it 10 minutes later I couldn't see anything
wrong or abnormal.  Traffic was low, no errors etc.

  I know two things it's not.

1) It's not mcmonitor's network, I checked it's packet drops and it
hadn't dropped any since sometime before noon Thursday.

2) It's not whatever was causing the RPF (rendezvous point) multicast
routing via OSPF errors (seen via 'show ip rpf events').   I'm relatively
certain those were caused by one of the things I fixed Thursday.  There
have been no more messages since I fixed it (1).

   It's possible we have two problems.  A significant one that occasional
caused repeated large scale drops which is fixed and a minor one that's
actually existed for some time but may be slightly worse.  If it continues
to behave the same, ie better but not fixed, then we'll probably upgrade
the IOS on the switch when I get back and if that gets rid of the less
frequent drops then good.

   It's asking a lot of the switches to *never* drop a multicast packet.
There is no guarantee they won't and it's not an error if they do.

   Trying to figure out why 3 packets out of 10's of thousands were
dropped 10-20 minutes ago is about as close to impossible as you can
get.  I'll keeping thinking about it but until we can figure out a way
to reproduce the problem in a controlled environment I don't have much
hope.

   I'll be in town today and tomorrow.  I'll be reachable via cell phone
now through Wednesday at 575-418-7368.  I'd still like to see one when
it happens.  I really need to see it as soon as possible,  within 2-3
minutes at least.


James

1) The switches had explicit rendezvous points set on the old routers.
With the routers out of the picture it was causing some confusion.  I
removed them late thursday when I realized they were still there.