[c-nsp] ME3600 IOS / SPAN

[Waris Sagheer (waris)]

Tassos,
Can you send me the list if SRs?
ME3800X should not be dropping packets with maximum queue-limit. We are
working on an enhancement to increase the queue-limit range.
Is there a SR for the counter issue?
Can you also send me the details of the last issue "matching of non
default classes"? I would like to do further investigation on that
topic.
I would recommend moving to 15.1(2)EY1 once it'll be available on CCO in
couple of weeks.

-Waris


-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Tassos
Chatzithomaoglou
Sent: Tuesday, November 22, 2011 2:06 PM
To: Nick Hilliard
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] ME3600 IOS / SPAN

We migrated from ME-3400 to ME-3800X and we noticed that we started
getting output drops 
on 1G interfaces, while the traffic was ~700 Mbps.
I know about bursts and so on, but the same traffic wasn't causing any
drops on the older 
platform.
We tried some output shaping service policies under the interface, but
the drops were 
still there, and then on the policy-map too.
We also noticed that the rate counters on the policy-map were totally
wrong.
Then we applied shaping service policies under the service instances and
increased the 
queue-limit to the max available, and the drops got very low (but still
existent).
Tac concluded that this is due to the default egress buffers being too
small, and we're 
waiting for the developers to come up with a solution.
Another issue we met, is that any match of non-default classes under the
service policies 
under the service instances wasn't working at all, when there was no
ingress L2 control 
traffic from the other side. I know it sounds very strange, but this was
the conclusion we 
came to, after doing many different tests.

To summarize, we have quite a few of strange cases with tac on this
platform.

--
Tassos


Nick Hilliard wrote on 22/11/2011 22:29:
> On 22/11/2011 18:44, Tassos Chatzithomaoglou wrote:
>> Unfortunately there are still a lot of features missing + some things
that
>> don't work well (primarily the -hardcoded- small egress buffers)
> Can you elaborate on this?
>
> Nick
>
_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/