Preview Tool

Cisco Bug: CSCvu08096 - NCS4k:6.5.26: xc traffic drop after doing RPFO twice then xc flap, due to pwe_lif failing to replay

Last Modified

Jun 19, 2020

Products (4)

  • Cisco Carrier Routing System
  • Cisco NCS 4016
  • Cisco IOS XR Software
  • Cisco NCS 4009

Known Affected Releases

6.5.26.BASE 6.5.31.BASE

Description (partial)

Xconnect over bundle traffic getting dropped after triggering one or two RPFO'ss followed by neighbor node reload or xconnect service flap.

Root cause:
The issue is because of PW object’s pw_id attribute failing to get replayed properly into the GRID during the case of RPFO. During RPFO, as part HADB replay, all the HW-IDs are instantly sent to the newly forming GRID for its reconstruction.
This replay feature works seamlessly across all the objects but for the case of PW object the IDs were mistakenly sent with extra 0x80 prefix and was getting rejected by the GRID. 
However, this issue wasn’t causing a problem in most instances because when the s/by node comes up, it requests GRID for the IDs and since they were never replayed in first place, they got reserved by these new requests from the s/by LCVM. 
But the catch here was, when the circuits come up in different order, it may end up getting different PW-IDs (as the GRID provides IDs on FCFS basis).. and clash with its own programming which is happened on issue node.

Xconnect service over bundle (on R1).
RPFO on router (R1).
Xconnect service flap or neighbor router (R2) reload or any condition which can flap the xconnect state.
Bug details contain sensitive information and therefore require a account to be viewed.

Bug Details Include

  • Full Description (including symptoms, conditions and workarounds)
  • Status
  • Severity
  • Known Fixed Releases
  • Related Community Discussions
  • Number of Related Support Cases
Bug information is viewable for customers and partners who have a service contract. Registered users can view up to 200 bugs per month without a service contract.