[j-nsp] MX80 Input Scheduling/Shaping

[Eric Van Tol]

Posting here for posterity.

The issue turned out to be that ingress scheduling can only be enabled on one MIC in the MX80. I had it enabled on both MICs and once I removed 1/0 and 1/1, the errors seem to have gone away. I only discovered this information in the Juniper MX Series, 2nd Edition book by Doug Hanks, Harry Reynolds and David Roy. If only the Juniper docs would mention this fact somewhere, or at least somewhere more obvious.

-evt

-----Original Message-----
From: juniper-nsp <juniper-nsp-bounces at puck.nether.net> On Behalf Of Eric Van Tol
Sent: Monday, October 8, 2018 7:06 AM
To: Tim Jackson <jackson.tim at gmail.com>; Alexander Arseniev <arseniev at btinternet.com>
Cc: jnsp <juniper-nsp at puck.nether.net>
Subject: Re: [j-nsp] MX80 Input Scheduling/Shaping

>Looks like you get 12 queues per MX80/104 for ingress shaping. Doesn't seem to be tied to QX at all. Egress you get per unit on the MIC slots though.
>
>https://www.juniper.net/documentation/en_US/junos/topics/task/configura
>tion/cos-configuring-ingress-hierarchical-cos-on-trio-mpc-mic.html
>
>https://www.juniper.net/documentation/en_US/junos/topics/concept/cos-hi
>erarchical-scheduling-understanding-fine-grained-queuing-for.html

So according to this, ingress scheduling is supported on the MX80 MIC slots. My question remains, which is, wtf is up with the inconsistent commit error? I checked PRs, but I've yet to find anything relevant.

-evt

On Fri, Oct 5, 2018, 8:45 PM Alexander Arseniev via juniper-nsp < juniper-nsp at puck.nether.net> wrote:

> Hello,
> Egress scheduling/shaping on MIC ports - correct, that's why I said 
> "roughly" equal.
> Ingress scheduling/shaping  requires Q or EQ MPC which is not 
> supported on MX80.
> Thanks
> Alex
>
> ------ Original Message ------
> From: sthaug at nethelp.no
> To: arseniev at btinternet.com; juniper-nsp at puck.nether.net
> Sent: 05/10/2018 20:20:06
> Subject: Re: [j-nsp] MX80 Input Scheduling/Shaping
>
> >>Ingress scheduling is supported only on Q and EQ MPCs - Juniper MX 
> >>series book, 2nd ed, page 598.
> >>
> >>MX80 COS capabilities are roughly equal to MPC1, without Q.
> >
> >Disagree. MX80 does per-VLAN queuing/scheduling/shaping just fine for 
> >ports on MICs. *Not* for the builtin 10G ports.
> >
> >Steinar Haug, Nethelp consulting, sthaug at nethelp.no
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net 
> https://puck.nether.net/mailman/listinfo/juniper-nsp
>
_______________________________________________
juniper-nsp mailing list juniper-nsp at puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsp at puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp