The SRAM is typically fully used. The problem is likely that the
GEIP is configured with the minimal SRAM & DRAM, upgrading to what
you're normally put in a VIP2/50 in your world would be a good
starting point...
George
> From cisco-nsp-request@puck.nether.net Fri Jul 14 18:31:17 2000
> Resent-Date: Fri, 14 Jul 2000 18:32:11 -0400
> Received-Date: Fri, 14 Jul 2000 18:27:38 -0400
> From: "Rubens Kuhl Jr." <rkuhljr@uol.com.br>
> To: "Martin, Christian" <CMartin@mercury.balink.com>,
> "cisco-nsp" <cisco-nsp@puck.nether.net>
> Subject: RE: [nsp] Buffers settings on GEIP controller (7500 router)
> Date: Fri, 14 Jul 2000 19:28:38 -0300
> In-Reply-To: <4F8C08CC6E76D311AD1F00508B787249F8F005@HERMES>
> Resent-From: cisco-nsp@puck.nether.net
> X-Mailing-List: <cisco-nsp@puck.nether.net> archive/latest/3247
> X-Loop: cisco-nsp@puck.nether.net
> Precedence: list
> Resent-Sender: cisco-nsp-request@puck.nether.net
>
>
>
> Although I've never heard of per-VIP buffer settings, the VIPs on this
> router seems not to follow the global buffer configuration commands...
>
> I/O memory may be the issue. The GEIP 'show mem' shows that the 8Mb SRAM is
> fully used:
> VIP-Slot4>show mem
> Head Total(b) Used(b) Free(b) Lowest(b) Largest(b)
> Processor 6068F1A0 127340704 25912004 101428700 101128988
> 101112860
> PCI 30000000 8388608 8382396 6212 6212 3988
>
> RSP has less fast memory, but it's not fully used:
> router# show mem
> Head Total(b) Used(b) Free(b) Lowest(b) Largest(b)
> Processor 61235EA0 249340256 22513572 226826684 224872320
> 224663032
> Fast 61215EA0 131072 93560 37512 37512 37468
>
> Data rate across Cybus is high (70%) but seems to be holding nice.
>
>
> Rubens Kuhl Jr.
>
>
> -----Original Message-----
> From: Martin, Christian [mailto:CMartin@mercury.balink.com]
> Sent: sexta-feira, 14 de julho de 2000 18:02
> To: 'Rubens Kuhl Jr.'; cisco-nsp
> Subject: RE: [nsp] Buffers settings on GEIP controller (7500 router)
>
>
> Rubens,
>
> You are looking at global particle pool information. You can thus increase
> the global data. I don't know that you can do it on a per-VIP basis.
>
> You can do something like this:
>
> buffers big permanent 1000
> buffers big max-free 1500
> buffers big min-free 500
>
> Note that you will need to be sure that you have enough I/O (MEMD) memory to
> do this. What is the data rate across the CyBUS? 'show contr cbus util'
> should reveal some things.
>
> regards,
> chris
>
> > -----Original Message-----
> > From: Rubens Kuhl Jr. [mailto:rkuhljr@uol.com.br]
> > Sent: Friday, July 14, 2000 4:26 PM
> > To: cisco-nsp
> > Subject: [nsp] Buffers settings on GEIP controller (7500 router)
> >
> >
> >
> >
> > I'm noticing a very large buffer miss rate on a GEIP controller (very
> > similar to a VIP2-50 with a Gigabit port adapter):
> >
> > Public buffer pools:
> > Small buffers, 104 bytes (total 25, permanent 25):
> > 25 in free list (20 min, 60 max allowed)
> > 2 hits, 0 misses, 0 trims, 0 created
> > 0 failures (0 no memory)
> > Middle buffers, 600 bytes (total 15, permanent 15):
> > 15 in free list (10 min, 30 max allowed)
> > 8 hits, 0 misses, 0 trims, 0 created
> > 0 failures (0 no memory)
> > Big buffers, 1524 bytes (total 100, permanent 5):
> > 6 in free list (5 min, 10 max allowed)
> > 2547640 hits, 861395 misses, 2231939 trims, 2232034 created
> > 104574 failures (0 no memory)
> > VeryBig buffers, 4520 bytes (total 148, permanent 50):
> > 148 in free list (40 min, 300 max allowed)
> > 104574 hits, 112 misses, 21 trims, 119 created
> > 0 failures (0 no memory)
> > Large buffers, 5024 bytes (total 0, permanent 0):
> > 0 in free list (0 min, 5 max allowed)
> > 0 hits, 0 misses, 0 trims, 0 created
> > 0 failures (0 no memory)
> > Huge buffers, 18024 bytes (total 1, permanent 1):
> > 1 in free list (0 min, 2 max allowed)
> > 0 hits, 0 misses, 0 trims, 0 created
> > 0 failures (0 no memory)
> >
> > The buffer settings are very different from the RSP, and I
> > think that's not
> > only because of particle pool memory allocation that the VIP
> > does. How to
> > set buffer parameters for VIP/GEIP ?
> >
> >
> > Rubens Kuhl Jr.
> >
>
>
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:14 EDT