lclark AT US.IBM DOT COM wrote:

> It is supposed to happen automatically, although not, perhaps, immediately.
> It is netmon's job to notice when there has been a change in the smartset,
> and then re-distribute the information to the appropriate mlm(s) on its
> next round.
> I believe that cycle is the same as the normal status polling cycle (eg 5
> min)
> when it checks the status of the mlms themselves, but I'm not positive.
> A stop/start of netmon ought to do it for the impatient.  I've seen it
> work.
>
> Anybody else?

It should happen automatically.  You can do a demand poll of the MLM to force
it along.

If it's not working call support. There are a few things that can cause this
not to work.  One is that netmon doesn't make a connection with nvcold.  In
V5.1, the way this used to work - and I don't know if it still works that way
in Version 6 -  is when netmon polled itself, it then tried to connect to the
nvcold daemon.  If you are running netmon with tracing turned on, you would
see the following in netmon.trace when netmon connects to the collection
deamon.

"collectiond_sock_fd = XX"

You'll have to stop netmon and start it with the tracing on, because once it
connects, you won't see this message again.


--
Ray Schafer                   | schafer AT tkg DOT com
The Kernel Group              | Distributed Systems Management
http://www.tkg.com