Datasheet

142
Release Notes for Cisco IOS Release 12.1E on the Catalyst 6500 and Cisco 7600 Supervisor Engine and MSFC
OL-2310-11
Caveats
• You might see SCHED-3-THRASHING messages when an SSH client sends text at a high input rate.
This problem is resolved in Release 12.1(26)E4. (CSCsa92622)
• A reload or spurious memory access occurs when HSRP rapidly changes states or goes up and down.
This problem occurs with IOS SLB configured and with virtual servers that are monitoring these
HSRP groups and probes that are configured on their server farms. This problem is resolved in
Release 12.1(26)E4. (CSCin94752)
• The identification field in all TACACS+ packets is always 0 when the synchronize (SYN) flag is set
and the TACACS+ packet goes through a filewall to the AAA server. The firewall interprets this 0
identification field as a Fragment Overlap Attack and drops additional new connections. This
problem is resolved in Release 12.1(26)E4. (CSCeh48684)
• With both WCCP and NDE configured, you might see numerous tracebacks caused by alignment
errors and CPU utilization might be unacceptably high. This problem is resolved in
Release 12.1(26)E4. (CSCsb21972)
• With a Supervisor Engine 1, Cisco IOS firewall load-balancing connection objects are not built for
non-TCP flows. This problem is resolved in Release 12.1(26)E4. (CSCsa88515)
• DLSw circuits are established over the same peer connection when DLSw load balancing is
configured and when there are multiple peers that have the dlsw icanreach mac-address mac_addr
command enabled with the same remote MAC address for the mac_addr argument. This problem is
resolved in Release 12.1(26)E4. (CSCsa45750)
• If RIP on a CE router redistributes a default route into BGP on a PE router, and then that default
route is redistributed back into RIP from a second CE router, the default route is not marked as
poisoned or withdrawn on the first CE router when the link between the first CE router and the PE
router is disconnected. This problem is resolved in Release 12.1(26)E4. (CSCeh06778)
• If two OSPF routing areas generate the same link-state advertisement (LSA) for a route and the route
is known on the Area Border Router (ABR) as an intra-area route, a summary LSA might not be
generated on the ABR if the route goes up and down. This problem is resolved in
Release 12.1(26)E4. (CSCeg62496)
• With a Supervisor Engine 1, in rare circumstances, a reload might occur if you enter the show mls
ip command. This problem is resolved in Release 12.1(26)E4. (CSCed00245)
• Erroneous “notPresent Temperature StatusValue Value =0” SNMP traps might be generated. This
problem is resolved in Release 12.1(26)E4. (CSCsa91816)
• When a Web Cache Communication Protocol (WCCP) service is enabled, and mask assignment is
configured as the assignment method, and five or more caches are in the service group, then protocol
messages sent to the cache may overflow and cause memory corruption and a reload. This problem
is resolved in Release 12.1(26)E4. (CSCeh56916)
• In rare situations a reload might occur, when there is a mix of Link State Advertisements (LSAs)
that travel throughout the Autonomous System (Types 5 and 11) and LSAs that travel within a
particular open shortest path first (OSPF) area (Types 1, 2, 3, 4, 6, 7, 9 and 10). This problem is
resolved in Release 12.1(26)E4. (CSCef93215)
• Following a reload, an OSPF designated router (DR) might fail to regenerate the network link-state
advertisement (LSA) when there is a shutdown interface with the same interface address in the OSPF
area. This problem is resolved in Release 12.1(26)E4. (CSCee36721)
• If you configure aggressive OSPF hello timers and dead timers, then during periods of high CPU
utilization, OSPF packets are not processed, and OSPF declares that OSPF neighbors are inoperative
(“down”). This problem is resolved in Release 12.1(26)E4. (CSCec42160)