Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

(ASCEND) PRI Rolldown problem, help!!



I've been noticing something lately since we have been getting close to
capacity on our MAX 4000s (hopefully we'll have some new units in soon!)
with regard to the way our rolldown currently works and the
unpredictability of it.

First, let me explain how we have our environment set up.  We have 4 MAX
4000 servers with 4 PRIs each installed.  2 PRIs per box are used for
modem access (we have 6 8-modem cards per box) and the other 2 PRIs per
box are for ISDN access.  We are running stacking and OSPF on all of our
boxes.  

We have a rolldown configured (via Southwestern Bell) such that all of the
dialup PRIs are linked together in a hunt group of 8 PRIs that span all 4
MAXen. Similarly, we have another such hunt group of 8 PRIs for our ISDN
customers.

What's happening is that sometimes (often) one of our dedicated ISDN
customers rings in on the hunt group for ISDN.  They get a connection in
the appropriate hunt group.  Their second channel then dials in and gets a
connection in one of the -dialup- (modem) hunt groups.  This doesn't
always happen but it happens often enough for me to notice.

It seems to me that this may be a feature of Stacking.  Does the MAX tell
the P-50 (or whatever) what number to dial for the second PRI?  Does the
Telco?  (How does either the MAX or the Telco know where to send it??)

I would say that this almost HAS to be true.  Here's an example of output
from a 'sh isdn' in a customer's P-50:


NL: CALL REQUEST: 64K, #2217536
NL: CALL REQUEST: 64K, #4743655


Where did 4743655 come from??  It's not programmed into the P-50 anywhere!
:)  It should -always- dial 2217536 no matter what.  4743655 is a valid
phone number but it's not in the hunt group that 2217536 is.

This is disturbing because it is taking channels away from our dialup
modem customers.  Is there a way to tell the Ascend where to send the
second channel of ISDN calls?  Or is it a Bell problem?  In any event, is
there any substantial evidence that it could be either?

If there's away around this problem I'd like to know...  If there's not a
way around it I would also like to know.

Thanks for the help!

-Rob-


++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>


Follow-Ups: