|
IPTRACE (requires 312PTC.EXE to be installed)
========
+ No command line to reset IPTRACE.LOG file, which has a tendency to grow very large. Now availabl eusing:-
LOAD IPTRACE <ipaddress> NEWLOG
+ When IPTRACE is executed with <ip address> NORESOLVE, it still attempts to resolve the From and To IP addresses
TCPIP
=====
+ UDP broadcasts on multihomed box would be sent out the first bound IP interface, and not all IP interfaces
+ IP routing table showing bogus next hop IP address when learning routes from OSPF LS database
+ Forwarding database not in sync with OSPF LSP database entries when default route filtered
+ RIPII does not send the correct subnet mask across a router ie. no support for discontiguous subnet masking
+ Fixed fragmentation problem in forwarding packets - caused
+ Server lost TCPIP network (cannot ping any IP node) while performing
connect/disconnect stress test.
+ When using an ASBR to import statically assigned IP addresses into the OSPF database, the external type 2 advertisement would be integrated correctly into the LSP database but the resulting forwarding database would show an incorrect next hop field.
+ TCPNewSendData can take more than 80 milliseconds to send a data out to the wire.
+ UDP Broadcast packets not broadcasting when the destination address is 255.255.255.255.
+ Server abends when you bind an unnumbered WAN link before having a numbered interface bound
+ When adding a secondary IP address, TCPIP did not check if that address is already in use - now generates an ARP.
+ Fixes an issue with LANDC denial of service attacks where the packets being sent are UDP, and not TCP based. The server will remain at 100% utilisation.
+ Fixes an issue with GroupWise POA hanging because TCPIP dropping internal requests
+ fixes problem where bogus IP Address being sent as source address on Unnumbered PPP IP links if the LAN Mapping address parameter is not used
+ Fixes an issue where TCPIP was generating Invalid ECBs when transmitting replies to sites where the maximum routable mtu packet size is smaller than the local adapter maximum packet size.
+ When we receive a FIN when we are in any other state other than EST, FW_1, FW_2 state, we pass the packet to the application and the application does not know what to do with this FIN. We drop these packets now.
+ Fixed fragmentation problem in forwarding packets - caused Abend issues that were solved by reducing the max. physical packet size
+ Unlimited command line length - previous versions of the stack caused abends when more than 132 characters were typed at the server console and the buffers filled up
+ Timeout value changes in TCPCON cause to show local snmp statistics
+ TCPCON crash after changing SNMP Access Configuration to non-existant server.
+ Server abends when you have OSPF, PPPRNS, and IPFLT enabled in one box. (1. NWCSTOP 2. Unload IPFLT and 3. Unload TCPIP would cause this to happen)
IPFLT31
=======
+ Toggling the enable/disable field in FILTCFG does NOT switch the status of the filters defined
+ race condition prevents assigning filter function pointers at load time
+ An existing External OSPF route failed to work when the router is restarted or after the filter status was changed
+ BOOTP Requests are blocked on ALL interfaces when a filter is defined for a specific interface
+ IP Packet forwarding exceptions don't work on subsequent connections when interface groups are used on AVM's MPR
+ When filters are setup on servers running PPPRNS, the external route filters block the propagation of the PPPRNS route when any change is performed to the filters.cfg file
TCPCFG.NLM (MPR only)
====================
+ fix the stub-subnet without debugging symbols. I have fixed the error checking routine so that it permits stub-subnet.
+ It only displays a warning message when users configure both bindings with the same IP address and mask. The correct thing is that it should first display the warning message and then force the user to reconfigure that IP address.
|
