Mailing List Archive

TurboIron-X24 mixed port type LACP?
Is there a way to get a TurboIron to link-aggregate a mix of built-in
copper ports and SFP ports, i.e. to build a >4x1000Base-T LAG to another
device without entirely using copper SFPs? I tried doing this recently,
found it would not work, and the error message was not terribly helpful.

Trunk: Group (23, 24, 25, 26, 27, 28) failed to create by 802.3ad
link-aggregation module.

That same message would get logged as each port that was meant to be in
the LAG got link.

I wondered if setting 25-28 to 1gb and disabling auto-negotiation would
make a difference, but did not try this as I didn't have out of band
access to the "other device" at the time I was trying to get the LAG up.

----------------------------------------------------------------------
Jon Lewis, MCP :) | I route
| therefore you are
_________ http://www.lewis.org/~jlewis/pgp for PGP public key_________
_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp
Re: TurboIron-X24 mixed port type LACP? [ In reply to ]
On 12/08/2015 16:22, Jon Lewis wrote:
> Is there a way to get a TurboIron to link-aggregate a mix of built-in
> copper ports and SFP ports, i.e. to build a >4x1000Base-T LAG to another
> device without entirely using copper SFPs? I tried doing this recently,
> found it would not work, and the error message was not terribly helpful.
afair, the 4x1G copper ports are handled using a different asic to the 10G
ports, and this is the sort of thing that might cause problems with LAG
formation. In general you probably don't want to mix 1G and 10G ports on
this platform, as the internal cell path between 1G and 10G ports will go
into store-n-forward mode. The platform only has 2M shared buffers, which
doesn't go far in store-n-forward mode.

Nick