Hi,

Assuming all my servers have nxge or e1000g nics, and they are all
connected to a stack of switches which are LACP capable, is there any
reason not to phase out IPMP in favor of link aggregates?

Is the feature mature enough for production?
Are there any known issues?
Any zone related issues?

Thanks
Yonah

we've been using aggregates in production for well over a year without
issues.  that includes in zones both 'normal' and clustered (with and
without failover zones).

> Assuming all my servers have nxge or e1000g nics, and they are all
> connected to a stack of switches which are LACP capable, is there any
> reason not to phase out IPMP in favor of link aggregates?
>
> Is the feature mature enough for production?
> Are there any known issues?
> Any zone related issues?

FYI, there was a bug when using LACP groups with virtual switches on
control domains for LDOMs.

On 01/06/09 13:48, Yonah Russ wrote:
> Hi,
> 
> Assuming all my servers have nxge or e1000g nics, and they are all
> connected to a stack of switches which are LACP capable, is there any
> reason not to phase out IPMP in favor of link aggregates?
> 
> Is the feature mature enough for production?
> Are there any known issues?
> Any zone related issues?
> 
> Thanks
> Yonah
> 


I see customers using link aggregations and IPMP together quite often. 
Create your link aggregations, then put them in IPMP groups.  Then the 
aggrs themselves can be standbys to each other or both active.

Perhaps I'm mis understanding your question : 0