Guest

Preview Tool

Cisco Bug: CSCus02689 - PRM fails to send NP_RUNNING state after NP_FAST_RESET_END

Last Modified

May 13, 2018

Products (8)

  • Cisco ASR 9000 Series Aggregation Services Routers
  • Cisco ASR 9922 Router
  • Cisco IOS XR Software
  • Cisco ASR 9010 Router
  • Cisco ASR 9904 Router
  • Cisco ASR 9006 Router
  • Cisco ASR 9001 Router
  • Cisco ASR 9912 Router

Known Affected Releases

4.3.4.LC 5.3.1.BASE

Description (partial)

Symptom:
When an MPA is inserted into a typhoon line card, the NP state is not in RUNNING_STATE and the state is displayed incorrectly in the 'show prm server context loc ....' command:

This can also happen on a trident line card when a NP recovery action happens (rare). The NP will recover, but the state will be displayed incorrectly in the 'show prm server context loc ...' command:

RP/0/RSP0/CPU0:ironman#sh prm server context loc 0/0/cpu0
Tue Dec  8 21:22:44.368 UTC

                Node: 0/0/CPU0:
----------------------------------------------------------------
PRM Server Process Context : 

        Restarted Since Bringup : No 
        Thread Info :
                TID        Name           Channel       
                ---        ----           -------       
                 1       prm_main            /dev/prm_server
                 9      prm_stats   /dev/prmLwmPath/evm_stats
                10         prm_tm     /dev/prmLwmPath/evm_tm
                11       prm_tcam   /dev/prmLwmPath/evm_tcam
                12         prm_ss     /dev/prmLwmPath/evm_ss
                13    prm_show_np   /dev/prmLwmPath/evm_show_np
                14   prm_mac_plus   /dev/prmLwmPath/evm_mac_plus
                15       prm_poll   /dev/prmLwmPath/evm_poll
                16      prm_npdbg   /dev/prmLwmPath/evm_npdbg
                17     prm_srh_lp   /dev/prmLwmPath/evm_low_priorit
                18  prm_mac_cache   /dev/prmLwmPath/evm_mac_cache
        NP State :
                NP0 :      RUNNING
                NP1 :     IN_RESET   <====

The NP is not in reset; the NP is actually in a Good State - able to handle traffic with one caveat:

Because the NP state is not reflecting the good state, any process that happens to key off this state when the process is started or restarted that process will not configure the NP until it is in a good state - resulting in traffic matching that particular feature not being handled by the NP.  A true example of this is OIRing an MPA and then installing the mcast pie after the OIR is complete.  The mcast traffic will not flow.  The problem could occur for a pie, SMU or a process restart - where a process is dependent on the np_state.

Conditions:
Any time an MPA is OIR'ed - either physically or by soft-oir using a command.  This only occurs on a typhoon-type line card.  It does not occur on any tomahawk-type line card.

Trident type LC can see this condition when a NP recovery action (Fast Reset) was taken for other unrelated NP situations. Even a canceled/aborted Fast Reset will incorrectly report IN_RESET.
Bug details contain sensitive information and therefore require a Cisco.com 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.