Wow! That's interesting. I've never seen your exact situation but I've sure seen some modules do strange things...even caused by the shelf-controller (especially under heavy load doing multishelf) Have you done the old swap slot/card/chassis thing? When you say hang, do see the hdlc2ec bouncing? (down, reset, load, boot, post). How about your fatal-history log or syslog info? It's hard to believe that there is NOTHING in these logs after what you have described happens. Does it happen only when you are in the card (open n n )? Again, debugging on a heavily loaded multishelf set doing tunneling may have made the SCM burp on you, etc. Of course, if this happens even when the slot is closed and no debugging is on, that'd be a null issue. Are you running 9.0.9 for a specific reason or can you downgrade to 9.0.6? It's nice when you ask a bunch of questions and get even more back, eh? Just probin'. - David -----Original Message----- From: Øystein Homelien [mailto:oystein at homelien.no] Sent: Sunday, November 17, 2002 1:42 PM To: ascend-users at bungi.com Subject: [Ascend] (ASCEND) ascend tnt hdlc2ec cards _hanging_ temporarily Anyone had problems with something like that? We have several TNTs that we see the same type of problem on. Every once in a while, typically several times per day, hdlc2ec cards will hang. We discovered this the following way: telnet to tnt log in open 1 x -- open a hdlc2ec card hdlc -p -- start displaying statistics of hdlc controllers The hdlc -p command typically prints out one line every second or so. We notice that the card hangs since suddenly it stops printing hdlc statistics every second. When this happens, the "open" connection to that particular card is dead/non-responsive. After 10-20-30-40 seconds, the card will come back up, in that the connection once again responds to commands and users connected through the card will have packets flowing to and from them again. While the card is hung, no ip/ppp traffic is forwarded to the users connected on that card at all (we see this by pinging a user on the card while this happens), but when the card comes back up everything works again (except typically some users have hung up because they do not expect 10-40 seconds of total packet loss during their dial-up session). Our TNTs are running 9.0.9, and as far as I can see this problem also exists in 9.0.4. We have no idea when this problem started happening, as we haven't found anything indicating what is happening in the logs, we just noticed that users sometimes experienced packet loss, and l2tp sessions timing out because of it. The most notable features we use are: * radius w/"call-check", auth and acct * stac * hdlc + modems * l2tp * multichassis mpp Anyone ever seen this, possible solutions? I am posting to the list because our local Lucent branch is being pretty much non-responsive to the problems we are having. They are simply not interested. Does Lucent care about its TNT customers who have invested heavily in these boxes? I'm starting to wonder, so this is a cry for help. Our customers certainly can't accept 30-second dead connections now and then, so unless we can get this TNT problem fixed, we are probably going to have to use the TNTs as doorstops and heaters, and go with Ciscos. It's just a pity that we actually believed in Ascend's promises back then when we had to choose between Cisco and Ascend for dialup equipment. -- Oystein Homelien, CVO | oystein at powertech.no PowerTech Information Systems AS | http://www.powertech.no/ Nedre Slottsgate 5, N-0157 OSLO | tel: +47-23-010-010, fax: +47-2220-0333 ++ Ascend Users Mailing List ++ To unsubscribe: send unsubscribe to ascend-users-request at bungi.com Archives: http://www.nexial.com/mailinglists/