Re: [nsp] Fast Mem on an RSP4

From: Deepak Jain (deepak@ai.net)
Date: Thu Sep 09 1999 - 15:01:01 EDT


This is what a proc mem shows:

Total: 122839328, Used: 115773136, Free: 7066192
 PID TTY Allocated Freed Holding Getbufs Retbufs Process
   0 0 54416 304 24004228 0 0 *Init*
   0 0 664 5181788 664 0 0 *Sched*
   0 0 346595268 310830556 191744 23492008 928800 *Dead*
   1 2 1704 432 14052 0 0 Virtual Exec
   2 0 0 0 6780 0 0 Check heaps
   3 0 136 0 6916 0 0 Pool Manager
   4 0 340 340 6780 0 0 Timers
   5 0 152 0 3932 0 0 OIR Handler
   6 0 0 0 6780 0 0 IPC Zone Manager
   7 0 0 0 6780 0 0 IPC Realm Manager
   8 0 1392 708 7464 0 0 IPC Seat Manager
   9 0 5804004 44719852 1229720 0 0 ARP Input
  10 0 96 0 6876 0 0 SERIAL A'detect
  11 0 228 0 7008 0 0 Microcode Loader
  12 0 0 0 6780 0 0 ATM ILMI Input
  13 0 0 0 6780 0 0 ILMI Process
  14 0 300 0 7080 0 0 IP Crashinfo Input
 PID TTY Allocated Freed Holding Getbufs Retbufs Process
  15 0 96 0 6876 0 0 DSX3MIB ll handler
  16 0 0 0 3780 0 0 Slave Time
  17 0 96 0 6876 0 0 Slave IPC OIR
  18 0 0 0 3780 0 0 Chassis Daemon
  19 0 0 11020 6780 0 0 IPC CBus process
  20 0 96 0 6876 0 0 MIP Mailbox
  21 0 136 0 6916 0 0 vcq_proc
  22 0 96 0 6876 0 0 CT3 Mailbox
  23 0 96 0 6876 0 0 CE3 Mailbox
  24 0 0 0 6780 0 0 FBM Timer
  25 0 0 0 6780 0 0 FDDI FDX Timer
  26 0 0 0 6780 0 0 SMT input
  27 0 268 268 6780 0 0 ATMSIG Timer
  28 0 20436 0 6876 0 0 MDFS RP process
  29 0 244 0 7024 0 0 Probe Input
  30 0 96 0 6876 0 0 RARP Input
  31 0 890463676 3715276 1858312 0 0 IP Input
  32 0 1333860 1344728 8028 0 0 TCP Timer
  32 0 1333860 1344728 8028 0 0 TCP Timer
 PID TTY Allocated Freed Holding Getbufs Retbufs Process
  33 0 2509544 780 15696 0 0 TCP
Protocols
  35 0 1912 840 7852 0 0 BOOTP Server
  36 0 3024 236552 15132 0 0 IP Background
  37 0 0 793715100 6780 0 0 IP Cache Ager
  38 0 244 0 7024 0 0 SYSMGT Events
  39 0 96 0 6876 0 0 SNMP ConfCopyProc
  40 0 0 0 6780 0 0 cbus utilization
  41 0 136 0 6916 0 0 Critical Bkgnd
  42 0 14296 640 6876 0 0 Net Background
  43 0 364 268 12876 0 0 Logger
  44 0 89788 268 6780 0 0 TTY Background
  45 0 80 13661024 6780 0 0 Per-Second Jobs
  46 0 385256 71812 6780 0 0 Net Periodic
  47 0 192 0 6972 0 0 Net Input
  48 0 1288 7638216 6780 0 11148400 Per-minute Jobs
  49 0 348524 5168 344036 0 0 CEF Background
  50 0 0 0 6780 0 0 FR ARP Input
  51 0 2024 68432 6780 0 0 FR LMI Input
 55 0 405272 0 6780 0 0 TCP Listener
  56 0 0 0 6780 0 0 CEF Scanner
  57 0 842768360 842767724 13416 0 0 IP SNMP
  58 0 3814560 3814464 6876 0 0 SNMP Traps
  59 0 8852 336 8048 0 0 NTP
  60 0 357823016 211223348 87640568 0 0 BGP Router
  61 0 28929160 0 6864 0 0 BGP I/O
  62 0 23415184 137931208 13628 0 0 BGP Scanner

The BGP doesn't seem out of sorts for this router. I could turn off soft
reconfiguration and free up most of that 90MB of RAM. Here is its BGP
summary:

BGP table version is 5776429, main routing table version 5776429
63720 network entries and 573492 paths using 26953992 bytes of memory
65951 BGP path attribute entries using 5294260 bytes of memory
92865 BGP route-map cache entries using 1485840 bytes of memory
Dampening enabled. 25 history paths, 104 dampened paths
279219 received paths for inbound soft reconfiguration
BGP activity 272135/208415 prefixes, 15338564/14765072 paths
2970489 prefixes revised.

Its DRAM utilization seems pretty stable.

          Fast memory

 Address Bytes Prev. Next Ref PrevF NextF Alloc PC What
60AB9EE0 44 0 60AB9F38 0 0 0 0 (fragment)
60AB9F38 5428 60AB9EE0 60ABB498 1 6016DD98 *Hardware IDB*
60ABB498 5428 60AB9F38 60ABC9F8 1 6016DD98 *Hardware IDB*
60ABC9F8 5428 60ABB498 60ABDF58 1 6016DD98 *Hardware IDB*
60ABDF58 5428 60ABC9F8 60ABF4B8 1 6016DD98 *Hardware IDB*
60ABF4B8 5428 60ABDF58 60AC0A18 1 6016DD98 *Hardware IDB*
60AC0A18 5428 60ABF4B8 60AC1F78 1 6016DD98 *Hardware IDB*
60AC1F78 5428 60AC0A18 60AC34D8 1 6016DD98 *Hardware IDB*
60AC34D8 5428 60AC1F78 60AC4A38 1 6016DD98 *Hardware IDB*
60AC4A38 5428 60AC34D8 60AC5F98 1 6016DD98 *Hardware IDB*
60AC5F98 5428 60AC4A38 60AC74F8 1 6016DD98 *Hardware IDB*
60AC74F8 5428 60AC5F98 60AC8A58 1 6016DD98 *Hardware IDB*
60AC8A58 5428 60AC74F8 60AC9FB8 1 6016DD98 *Hardware IDB*
60AC9FB8 5428 60AC8A58 60ACB518 1 6016DD98 *Hardware IDB*
60ACB518 5428 60AC9FB8 60ACCA78 1 6016DD98 *Hardware IDB*
60ACCA78 5428 60ACB518 60ACDFD8 1 6016DD98 *Hardware IDB*
60ACDFD8 5428 60ACCA78 60ACF538 1 6016DD98 *Hardware IDB*
60ACF538 5428 60ACDFD8 60AD0A98 1 6016DD98 *Hardware IDB*
60AD0A98 5428 60ACF538 60AD1FF8 1 6016DD98 *Hardware IDB*
60AD1FF8 5428 60AD0A98 60AD3558 1 6016DD98 *Hardware IDB*
60AD3558 5428 60AD1FF8 60AD4AB8 1 6016DD98 *Hardware IDB*
60AD4AB8 5428 60AD3558 60AD6018 1 6016DD98 *Hardware IDB*
60AD6018 5428 60AD4AB8 60AD7578 1 6016DD98 *Hardware IDB*
60AD7578 5428 60AD6018 60AD8AD8 1 6016DD98 *Hardware IDB*
60AD8AD8 5084 60AD7578 0 0 0 0 6015FA8C (fragment)

I can't make heads or tails of the fast memory utilization which is why I
brought this up.

Thanks,

Deepak Jain
AiNET

On Thu, 9 Sep 1999, Sean Butler wrote:

>
> Try "show proc mem" and just a plain "show mem" looking at
> the "what" column.
>
> That should give you a pretty good idea of what's holding the
> memory or if its just lost...
>
> a CEF table on a router with a full Internet routing table should
> only be about 15M. 128M on router *should* be plenty to do
> just about anything you want.
>
> What version are you running?
>
> /Sean
> ___________________________________
> Sean Butler, CCIE #3897
> AT&T Global Services -- OpenNet Support
> Phone: 727-533-8830
> Fax: 813-878-5475
>
>



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:05 EDT