I am having a problem with one of my TCH chassis, out of a total of four. The chassis in question is running a hiperarc and two dsp's connected to PRI lines primarily servicing ISDN. The other three are hiperarc w/ 12 quads and two dsp's on T1's. All are running the same level of code. The problem is that the ISDN chassis is losing some of the accounting packets, mainly the stop packets. It appears to occur more frequently with certain users, but I think that is mainly because they log on and off a lot. I don't see this on any of the other chassis's at all. I have compensated for this by using a radius server that monitors the chassis via snmp to keep the active user list accurate. The problem is that if a client tries to log in before the snmp cycle occurs they fail because of too many active sessions. I have another hack for that but something is wrong that needs to be fixed. I have tried two different NT machines in different network configurations (switch vs. hub on same segment), and I have tried a Linux machine configured to receive the radius accounting packets. All have experienced the problem so I have concluded it is in the chassis. I was kinda hoping it was the NT machines, because it is easy to replace;) Anyway, I am lost on what to else look for. All the systems are configured in the same way or at least I think so. Is there a way to do a configuration dump from multiple chassis's and then run them through a comparison? What parts of the system are responsible for generating the stop message? One thing I haven't tried yet is swapping the hiperarcs between chassis's and see if the problem moves. Is this worth a shot or could it be the NMC. All of mine are real old. Thanks for any insight you can offer... Mark Thornton San Marcos Internet, Inc. 512-393-5300 - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.