[c-nsp] mitigate output drops on Cisco 2960G platform

[Gert Doering]

Hi,

On Fri, Sep 20, 2013 at 07:56:26PM +0300, Martin T wrote:
> I have three WS-C2960G-24TC-L switches which all have about ten
[..]
> Any suggestions other than reduce the traffic on switch uplink ports?

"Get some real switches"...

Not exactly constructive, sorry.  Been there, felt the pain, ranted on
cisco-nsp on 2960 and their lack of buffers, and worse, their lack of
documentation on their lack of buffers.  There's *lots* of it in the 
archives...

If you don't need QoS, turning it *off* with "no mls qos" will help
somewhat (as then all buffers are available for your packets, not only
1/4 which is allocated to that particular queue).  What we had to 
resort to is to make the uplink ports 2xGE or even 4xGE channels to
exceeding avoid a load of ~25-30 percent per link - we had fairly
bursty ingress traffic, which (for microseconds a time) overfilled
the buffers, while the average traffic across all ingress ports was
WAY below the rate the egress could handle.   Microbursts, too little
buffering, switches not suited for that traffic.

gert
-- 
USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             gert at greenie.muc.de
fax: +49-89-35655025                        gert at net.informatik.tu-muenchen.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 305 bytes
Desc: not available
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20130920/1e472b51/attachment.sig>