[j-nsp] scb RT: Failed prefix delete IPv4 - x.x.x.x/24 (No memory)

[Matt Yaklin]

Hi list,

I am seeing these error messages.

/kernel: RT_PFE: RT msg op 2 (PREFIX DELETE) failed, err 1 (Unknown)
scb RT: Failed prefix delete IPv4 - x.x.x.x
scb RT: Failed prefix delete IPv4:0 - x.x.x.x/24 (jt delete failed)/24 (No memory)

Besides failed prefix deletes, I am also seeing them for prefix adds.

I thought this error message was rather clear cut to what the problem
could be. As in the scb was running out of memory. But that does not
seem to be the case based on the output below.

This is an old M40 running an older JunOS release. A very old release.
5.7R2.4. (I expect a few chuckles here).

user at router> show chassis scb
SCB status:
   Temperature                 26 degrees C / 78 degrees F
   CPU utilization              1 percent
   Interrupt utilization        0 percent
   Heap utilization            28 percent
   Buffer utilization          44 percent
   Total CPU DRAM             128 MB
   Internet Processor I           Version 1, Foundry IBM, Part number 3
   Start time:                    2005-03-22 19:43:37 UTC
   Uptime:                        1650 days, 4 hours, 11 minutes, 5 seconds

user at router> show chassis routing-engine
Routing Engine status:
     Temperature                 31 degrees C / 87 degrees F
     DRAM                       512 MB
     Memory utilization          50 percent
     CPU utilization:
       User                       1 percent
       Background                 0 percent
       Kernel                     1 percent
       Interrupt                  0 percent
       Idle                      98 percent
     Model                          RE-1.0
     Start time                     2005-03-22 19:41:26 UTC
     Uptime                         1650 days, 4 hours, 15 minutes, 30 seconds
     Load averages:                 1 minute   5 minute  15 minute
                                        0.07       0.06       0.02


But after some reading old of posts to this mailing list I
saw this post:

http://puck.nether.net/pipermail/juniper-nsp/2008-October/011550.html

root at router> start shell pfe network scb

SCB platform (200/266Mhz PPC 603e processor, 128MB memory, 256KB flash)

SCB(router vty)# show jtree 0 memory
Memory Statistics:
     4194304 bytes total (4 banks)
     4194304 bytes used
           0 bytes free
        4064 pages total
        4064 pages used
           0 pages free
          31 max freelist size

Free Blocks:
  Size(b)    Total(b)        Free       TFree       Alloc
--------  ----------  ----------  ----------  ----------
        8     3102208         153           0      387623
       16     1092048         125           0       68128
       24          48           0           0           2
       32           0           0           0           0
       40           0           0           0           0
       48           0           0           0           0
       56           0           0           0           0
       64           0           0           0           0
       72           0           0           0           0
       80           0           0           0           0
       88           0           0           0           0
       96           0           0           0           0
      104           0           0           0           0
    Total     4194304


So I assume this is where I am out of memory? This box has been
up for a long time and my question is would a reboot allow it to
continue to function for many years to come or I being out of
memory now will simply happen again right away? This M40 has
served us well over the years.

I have a M10 I have been meaning to use to replace this which
is a nicer box hardware wise and I am trying to determine if this
box is no longer useful on our network.

Thank you for any assistance or advice.

matt at g4.net