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

RE: (ASCEND) 1.3Ap24 for the MAX TNT



Kevin,
	Why do the release notes say that Stac compression is turned off
on the HDLC card?  We have been running it since a few days after it was
released, and stac compression seems to be working fine.  I can get at
least 3 to 1 compression on a d/l of a text file from a server at my
office.   I asked my friend of mine that works for you guys about it and
he said he was not sure why the notes say it was disabled.

		Chris


=-=-=-=-=-=-=-=-=-=-=
Chris R. Fairbanks
Director of Internet Services

WinterLAN, Inc.
2910 Telegraph Ave.
Berkeley, CA 94705
Voice +1.510.486.1812
Fax   +1.510.486.1796
http://www.wli.net/



> -----Original Message-----
> From:	Kevin A. Smith [SMTP:kevin@ascend.com]
> Sent:	Tuesday, January 27, 1998 6:05 PM
> To:	Ascend User-Group
> Subject:	(ASCEND) 1.3Ap24 for the MAX TNT
> 
> A bit of old news for some, but just to be formal:
> 
> 1.3Ap24 for the MAX TNT has been given to support.  This release
> contains
> the following corrections and enhancements.
> 
> Corrections in 1.3Ap24:
> *  TR 1754: Old IP route entries were not cleared under some
> conditions.
> *  TR 2110: Connections/Session count was not accurate.
> *  TR 2257: Digital call routing could sometimes stop on MP+ calls.
> *  TR 2324: Shelf-controller reset with Warning 191 but no FE.
> *  TR 2383: Formatting both flash cards could sometimes cause a reset
> with
> FE 104.
> *  TR 2443: MAX TNT now supports up to 15 FrameLine cards.
> *  TR 2456: RADIUS authentication could fail after sustained
> dialin/dialout
> traffic.
> *  TR 2473: MAX TNT did not grab all outgoing routes and profiles from
> RADIUS.
> *  TR 2483: Modems reset with FE 2.
> *  TR 2502: ISDN users could not connect and Ifmgr output showed
> incorrect
> names.
> *  TR 2504: Second flash card could be unavailable.
> *  TR 2506: SNMP traps now report loss of communication to slave
> shelf.
> *  TR 2521: A write error occurred when loading code to flash.
> *  TR 2522 / 2822: FE 3001 on Ethernet cards is now reported correctly
> as a
> warning.
> *  TR 2537: MAX TNT sometimes sent packets to the wrong WAN interface.
> *  TR 2552: 56K- modem cards showed high CRC errors during JPEG file
> down
> load.
> *  TR 2571: Ascend MIB  did not report second slot of modem cards as
> Occupied.
> *  TR 2582: Calls nacked via DNIS did not return User Busy.
> *  TR 2619: Ethernet, HDLC, and modem cards could generate FE 2 under
> heavy
> load.
> *  TR 2625: System sometimes reset with FE 2 after several days of
> heavy load.
> *  TR 2628: RADIUS accounting records could replace names with modem
> numbers.
> *  TR 2629: MAX TNT could lose its configuration following a system
> crash
> or reset.
> *  TR 2630: V.32 modems had difficulty connecting to 56K modems due to
> V.8bis tone.
> *  TR 2647: V34 modems reset with FE2.
> *  TR 2648: RADIUS accounting could report incorrect information for
> async
> PPP calls.
> *  TR 2662: Abnormal tone before V.8 56K tone blocked some modem
> callers.
> *  TR 2666: After 7 to 24 hours, E1 channels using R2 signaling became
> inoperative.
> *  TR 2685: 56K modem cards showed low sync rate ("dead air") on
> heavily
> loaded systems.
> *  TR 2688: The shelf-controller could crash following a Warning 108
> when
> receiving a small bus *  packet from a slot card.
> *  TR 2696: Network Management option was not enabled with the update
> hash
> code.
> *  TR 2711: OSPF could cause FE 29 on the shelf-controller.
> *  TR 2714: HDLC card sometimes wouldn't load when unit is busy.
> *  TR 2715: HDLC cards appeared to fail with Timeout Failure error
> messages.
> *  TR 2718: Ethernet slot card crashed when Proxy ARP was enabled.
> *  TR 2720: ISDN users could fail to connect due to erroneous RADIUS
> timeouts.
> *  TR 2723: Second channel was sometimes dropped on a two-channel
> permanent
> switched MP+ *  connection.
> *  TR 2732: FE 1s were sometimes observed on the shelf-controller.
> *  TR 2733: Ethernet card could sometimes generate a FE 8.
> *  TR 2735: Multishelf system couldn't load more than 16 slot card
> simultaneously.
> *  TR 2736: Modem card reported "WANSESS: newSession: Out of
> Sessions."
> *  TR 2739: FE 8 or FE 104 could occur during external authentication
> of
> PPP sessions.
> *  TR 2744: IP port caching enabled.
> *  TR 2745: Coredump fixes.
> *  TR 2758: K56 Modem cards could reset after a fatal error and not
> attempt
> to reload.
> *  TR 2768: TCP ISN generation is now more secure.
> *  TR 2798: When an MP call dropped a channel, the MAX TNT could stop
> routing to the far-end *  device under certain conditions.
> *  TR 2808: Userstat command sometimes did not report the actual
> number of
> users.
> *  TR 2834: Slot cards reset when trying to log Warning 104.
> *  TR 2841: Modem transmits level incorrect for K56 modems.
> *  TR 2843: End-of-packet character did not work properly for
> TCP-Clear calls.
> *  TR 2897: Invalid emergency log message seen on console.
> *  TR 2969: Call connect rate could degrade under heavy call volume.
> *  TR 2972: Encapsulation method was  not enforced in local Connection
> profile.
> *  TR 2983: Ethernet card could reset without generating a fatal log
> message.
> *  TR 3023: FE 36 was observed when MP calls spanned multiple HDLC
> cards.
> *  TR 250036: Security problem with CLID authentication.
> *  TR 250050: System Up diagnostic was not logged after a reset.
> *  TR 250071: RADIUS accounting packet octet counts were inaccurate.
> *  TR 250075: The system did not refresh RADIUS routes if
> dialout-poison
> was enabled.
> *  TR 250078: RADIUS address pools were not deallocated properly.
> *  TR 250080: The Expect Callback feature did not work.
> *  TR 250102: HDLC cards ran out of memory intermittently.
> *  TR 250108: Call rejected with more modems in use than callers.
> *  TR 250117: Call-Filters applied in a Connection profile did not
> work.
> *  TR 250127: CIDR routing caused low memory conditions.
> *  RADIUS initialization on slave-shelf 56K modem cards failed
> intermittently.
> *  N/A  HDLC cards sometimes operated slowly under light load.
> *  N/A  User names could be reported inaccurately in SNMP accounting
> events.
> *  N/A  TCP-clear sessions could generate extra SNMP name or service
> change
> events.
> *  N/A  Writing a T1 or E1 profile with an invalid index could
> generate a
> FE 1.
> *  N/A  RADIUS accounting ON and OFF packets sometimes timed out.
> *  N/A  CLID and DNIS were not contained in some syslog messages.
> *  N/A  SNMP accounting progress codes and disconnect reasons were
> incorrect for immediate TCP *  sesN/A  sions.
> *  N/A  Ethernet card failed to come up under certain conditions.
> *  N/A  PPP and immediate TCP clear sessions invoked from the terminal
> server now use memory *  more N/A  efficiently.
> *  N/A  ICMP redirects were not sent in some instances.
> *  N/A  VJ compression modem sessions now function with a smaller
> number of
> slots.
> *  N/A  Fatal Error 102 sometimes occurred on 56K-modem cards.
> *  N/A  Out-of-memory error on the shelf-controller has been fixed.
> *  N/A  HDLC cards could occasionally crash under heavy traffic.
> *  N/A  T-Online: The MAX TNT can now dialout MP+ calls when T-Online
> is
> enabled.
> *  N/A  T-Online: Calling DTPT Connection profile made LAN
> inaccessible.
> *  N/A  T-Online: Problem 503 has been fixed.
> *  N/A  T-Online: Problem 30965 has been fixed.
> *  N/A  Potential memory leaks were fixed.
> *  N/A  Routes learned from RADIUS showed up twice in the routing
> table.
> *  N/A  VJ compression could cause reduced throughput.
> *  N/A  HDLC card failed to come up about 15% of the time.
> *  N/A  RADIUS packet type 51 was not delivered from Ethernet cards.
> *  N/A  IP pool addresses became unavailable when used for MP or MP+
> calls.
> *  N/A  A T3 card provisioned with PRI DS1s could crash.
> *  N/A  Coredump was not supported on local Ethernet interfaces.
> *  N/A  Slot cards in DIAG state remained UP if coredump failed.
> *  N/A  HDLC card FE8 under heavy MP+ and MP traffic and dynamic
> addressing.
> *  N/A  HDLC card could crash if a Ping packet arrived during call
> setup.
> *  N/A  T1 interfaces could come up as unavailable.
> *  N/A  Users were not disconnected via SNMP request.
> *  N/A  HDLC cards could crash receiving LCP frame from a slave MP or
> MP+
> channel.
> *  N/A  Outbound calls used the first DSL regardless of a line's
> state.
> *  N/A  Host cards could generate a FE 8 under heavy MP+ load.
> *  N/A  MAX TNT could stop accepting calls.
> *  N/A  Data transfer to client over SLIP/CSLIP sessions could hang.
> *  N/A  Shelf-controller could crash when a T3 card went down with a
> fatal
> error.
> *  N/A  HDLC cards in slave shelves could sometimes fail to load.
> *  N/A  NFAS fixes.
> *  N/A  Outgoing MPP calls for a second channel could sometimes fail
> authentication.
> *  N/A  Slot card profile requests could time out on very busy
> multishelf
> systems.
> *  N/A  Slave shelves didn't log RESET to master's fatal log.
> *  N/A  All channels of MP and MP+ connections were sometimes not
> cleared.
> *  N/A  Data corruption could occur if a host card when down in a
> multishelf system.
> *  N/A  RIP could send duplicate packets to the multicast/broadcast
> address on Ethernet.
> *  N/A  IGMP query packets could keep  an idle connection alive.
> *  N/A  Increases in memory efficiency for slot-card route caches.
> *  N/A  Dynamic bandwidth allocation (DBA) sometimes did not work.
> *  N/A  Slot cards could time out or generate FE2 under high traffic
> conditions.
> *  N/A  The IPcache cache command did not display one page of output
> at a
> time.
> *  N/A  Modem card  FE 1 following a Warning 106 or 104.
> *  N/A  T3 card could sometimes crash the shelf when dumping core.
> *  N/A  HDLC card could time out when trying to load many RADIUS
> routes
> after a reset.
> *  N/A  RADIUS external routes could cause host card memory leak.
> *  N/A  Warning 191 & 192 and other TCP fixes for TCP-clear sessions.
> *  N/A  Incoming PPP session brings up several DTPT sessions.
> *  N/A  RADIUS  accounting  no longer records  DTPT sessions.
> *  N/A  Interface could be stuck in Pending state if dialout call
> failed.
> *  N/A  HDLC and modem cards could periodically reset.
> *  N/A  Excessive CRC errors on some multishelf systems.
> *  N/A  100Base-T port  could lose sync with a GRF switch on Ethernet.
> *  N/A  Client-Default-Gateway should not allow a netmask
> specification.
> *  N/A  Memory leak fixed for incoming modem calls.
> *  N/A  100Base-T interface was displayed in UP state when card is
> downed
> or removed.
> *  N/A  Changes to Netstat command output
> *  N/A  Slot cards could take several minutes to come up in a loaded
> multishelf system.
> *  N/A  Change in router behavior when buffers are full
> *  N/A  TCP-Clear traffic on a slot card did not use the IP route
> cache.
> *  N/A  NasPortID (5) and NasPortType (61) were reported in Accounting
> On/Off packets.
> *  N/A  Modem slot card FE2 with RADIUS authentication.
> *  N/A  Warning 140 observed under heavy load.
> *  N/A  Ethernet controller could overwrite current packet.
> *  N/A  System did not respond with message when second Setup message
> was
> received.
> *  N/A  RADIUS authentication did not work unless the system address
> was set.
> *  N/A  Intermittent problems with digital connections.
> *  N/A  RIP could sometimes send unnecessary updates.
> *  N/A  FE 175 amd FE 8 was observed with heavy TCP-Clear traffic.
> *  N/A  FE 179 was observed related to UDP event queues.
> *  N/A  Could not establish SLIP connection using IP Pools.
> *  N/A  Phantom interfaces prevent subsequent calls.
> 
> Enhancements in 1.3Ap24:
> *  Support for Rockwell K56flex code version 1.160T
> *  Distinct ID sequences for RADIUS authentication and accounting
> *  Unique RADIUS accounting IDs based on source port number
> *  Command for displaying WAN session data
> *  Support for 64K and 56K calls over channels using R2 signaling
> *  Slot card and system uptime information
> *  Terminal server login timeout
> *  Fatal crash information on console
> *  Multiple NFAS groups on T1 and T3 cards
> *  MP and MP+ bundled channels span HDLC cards and shelves
> *  IP port caching
> *  ATMP attributes in RADIUS accounting Stop records
> *  New disconnect code (210-slot card down)
> *  Finger (RFC 1288) support and Userstat enhancements
> *  Inactivity timers for ATMP tunnels
> *  TCP-timeout parameter enabled
> *  New Local-Profiles-First setting
> *  Userstat -k to terminate user sessions
> *  New call-routing sort method for digital calls
> *  IP-Pools command
> *  Configurable cause element in ISDN Disconnect packets
> *  Separate transmit and receive data rates reported
> *  Shelf-controller proxy RADIUS accounting
> *  V.34 setting for 56k modem cards
> *  Local DNS table
> *  Turn on multishelf trap by default
> *  DTPT sessions to the ZGR identified in Session MIB
> *  Performance enhancements for TCP-Clear calls
> 
> Known problems in 1.3Ap24:
> * To route heavy OSPF traffic on large networks, the MAX TNT may
> require
> the additional memory provided by the 32MB JDEC DRAM card (Model
> TNT-SP-DRAM32, Part Number 0800-0489-001).
> * In this release, Stac compression support on HDLC cards is turned
> off. 
> * In this release, output of the OSPF Database diagnostic command does
> not
> display link state entries of type ASE (Type-5). However, as a
> workaround,
> administrators can still view the Type-5 link states by using the
> following
> form of this command: admin> ospf database ase 
> 
> -- Patti
> 
> 
> Kevin
> 
> 
> ++ Ascend Users Mailing List ++
> To unsubscribe:	send unsubscribe to
> ascend-users-request@bungi.com
> To get FAQ'd:	<http://www.nealis.net/ascend/faq>
++ 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: