[j-nsp] M20 PFE/SSB-E problems

[Josef Buchsteiner]

Jonas,
      you had the same problem in November 2004 and I advised to
      monitor the 'show nhdb zone' command to see the utilisation of
      the next-hops you have due to arp entries. Have you done this ?

      you had one time at least 14000 entries and you have
      non-enhanced FPC which can store about 17k. Once you have an
      aggregate interface and all two links are on the same FPC you
      could then only have 8,5k since you needs a next-hop for each
      member link so you can do the math. depending on your setup.

      Did you had network troubles ? From you message back in November
      you indicated a number of 3600 was usual. Is someone doing test
      on a network and creating arp entries ?

      Josef

Saturday, March 19, 2005, 9:56:47 PM, you wrote:

  
JF> Hi,

JF>  suddenly i am getting the following errors on a M20:
JF>  Mar 19 18:48:54  ffm2-edge /kernel: RT_PFE: NH IPC op 31 (CHANGE
JF>  AGGREGATE NEXTHOP) failed, err 6 (No Memory)
JF>  Mar 19 18:49:12  ffm2-edge last message repeated 2 times
JF>  Mar 19 18:49:17  ffm2-edge /kernel: RT_PFE: NH IPC op 31 (CHANGE
JF>  AGGREGATE NEXTHOP) failed, err 6 (No Memory)
JF>  Mar 19 19:01:14  ffm2-edge last message repeated 42 times
JF>  Mar 19 19:09:47  ffm2-edge last message repeated 39 times
JF>  Mar 19 19:19:39  ffm2-edge last message repeated 10 times
JF>  Mar 19 19:48:48  ffm2-edge /kernel: RT_PFE: NH IPC op 29 (ADD AGGREGATE
JF>  NEXTHOP) failed, err 6 (No Memory)
JF>  Mar 19 19:48:48  ffm2-edge /kernel: RT_PFE: NH IPC op 31 (CHANGE
JF>  AGGREGATE NEXTHOP) failed, err 6 (No Memory)
JF>  Mar 19 20:02:41  ffm2-edge /kernel: RT_PFE: RT msg op 1 (PREFIX ADD)
JF>  failed, err 1 (Unknown)

JF>  [messages are repeating over and over]

JF>  then sometime later:
JF>  Mar 19 20:02:41  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:41  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss, nh 11716, ifidx 68, nhifl 0, flag 0x0, cos 0
JF>  Mar 19 20:02:41  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:41  ffm2-edge /kernel: RT_PFE: RT msg op 1 (PREFIX ADD)
JF>  failed, err 1 (Unknown)
JF>  Mar 19 20:02:41  ffm2-edge last message repeated 7 times
JF>  Mar 19 20:02:41  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss, nh 11814, ifidx 76, nhifl 0, flag 0x0, cos 0
JF>  Mar 19 20:02:41  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:42  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss, nh 14090, ifidx 68, nhifl 0, flag 0x0, cos 0
JF>  Mar 19 20:02:42  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:42  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss, nh 14091, ifidx 68, nhifl 0, flag 0x0, cos 0


JF>  [repeating, beeing some.ip.addre.ss an advertised ip address inside our
JF>  ip space]

JF>  after that:
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix add IPv4:0 -
JF>  some.ip.addre.ss (radix add failed)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)
JF>  Mar 19 20:02:53  ffm2-edge ssb RT: Failed prefix delete IPv4:0 -
JF>  some.ip.addre.ss (unknown prefix)

JF>  [repeating over and over]

JF>  and after this again the "change aggregate nexthop" error.

JF>  The SSB looks fine:

JF>  SSB status:
JF>  Slot 0 information:
JF>    State                                 Master   
JF>    Temperature                        35 degrees C / 95 degrees F
JF>    CPU utilization                    21 percent
JF>    Interrupt utilization               1 percent
JF>    Heap utilization                   64 percent
JF>    Buffer utilization                 58 percent
JF>    Total CPU DRAM                     64 MB


JF>  As well as the RE:

JF>    Slot 0:
JF>      Current state                  Master
JF>      Election priority              Master
JF>      Temperature                 25 degrees C / 77 degrees F
JF>      CPU temperature             27 degrees C / 80 degrees F
JF>      DRAM                       768 MB
JF>      Memory utilization          42 percent
JF>     CPU utilization:
JF>        User                       2 percent
JF>        Background                 0 percent
JF>        Kernel                     3 percent
JF>        Interrupt                  1 percent
JF>        Idle                      94 percent


JF>  Running 7.0R2.7

JF>  Regarding operational impact this error causes some IP addresses not
JF>  beeing able to be reached from outside. These IP addresses can be
JF>  reached by the router, e.g. a simple ping works. However nothing from
JF>  outside reaches them.

JF>  Anyone seen this before?

JF>  Regards,
JF>  Jonas




JF>  _______________________________________________
JF>  juniper-nsp mailing list juniper-nsp at puck.nether.net
JF> http://puck.nether.net/mailman/listinfo/juniper-nsp