
394
Cross-Platform Release Notes for Cisco IOS Release 12.0S
OL-1617-14 Rev. Q0
Caveats
Resolved Caveats—Cisco IOS Release 12.0(33)S6
We observed checksum errors.
Workaround: After seeing checksum errors in the IP traffic statistics, we tried shutting the
core-facing interface in the olist. After that, the problem disappeared. When we added that interface
back, the problem was reproduced again. We suspect the following to cause this issue.
When we have core and VRF interfaces on the egress LC (E5), the PIM packet has to be forwarded
on the core-facing interface and also has to be punted to the RP. In the E5, this is done by recycling
the packet. In the first cycle, the packet will be sent to the core interface; in the second cycle, the
packet will be decapsulated and punted to the LC CPU.
Only the head gets recycled for different passes. The tail will be stored in the stingray. When the
packet is punted to the LC CPU, the LC CPU will copy the tail from the stingray, attach it to the
head, and send it to the RP. We suspect that this copy is not happening properly and the RP is seeing
PIM checksum errors.
• CSCta37296
Symptoms: On a Cisco 12000 ATM linecard, threshold drops are observed on a service policy that
is configured with WRED, and these drops occur even if no random drops are incrementing.
Conditions: This symptom is observed only when using WRED, and it is observed even under low
utilization of the service policy. Threshold drops will happen only for RP-generated packets even
though there are no random drops.
Workaround: To stop the drops, remove WRED and configure a standard queue limit.
Further Problem Description: Some of the packet types that are generated by the linecard (such as
ICMP echo response) will still undergo threshold drops.
• CSCta45402
Symptoms: In an MVPN setup with a CE connected via an MLPPP interface, auto- RP packets are
not being punted to the RP and the RP entry times out after 180 seconds.
Conditions: This symptom is observed either when a link flaps on a member of the MLPPP interface
or when output QoS is applied on the MLPPP interface.
Workaround:
1) RP# clear ip mroute vrf <vpn> 224.0.1.40
2) LC# clear ip mds all
3) Configure static RP.
4) Remove the output policy on the outgoing Multilink.
• CSCta58995
Symptoms: A Cisco 7200 PE is dropping *small* frames on an AToM FRoMPLS tunnel.
Conditions: This symptom is observed in an FR IP IW case when frames that are less than 60 bytes
are sent from a Cisco 12000 series router (PE on the other side).
Workaround: There is no workaround.
• CSCta69919
Symptoms: On a Cisco 12000 series router with ISE line cards and an IPv6 ACL, after a reload or
RP switchover, the ACL does not match traffic correctly.
Conditions: This applies to IPv6 ACL.
Workaround: Delete and recreate the ACL.
Comentarios a estos manuales