On 9/13/07, Peter Memishian <peter.memishian at sun.com> wrote: > I've once again updated the IPMP High-Level Design to cover some > additional implementation experience over the past month[1]: > > http://opensolaris.org/os/project/clearview/ipmp-highlevel-design.pdf > > Changes from the previous version: > > * Made semantically significant updates to the sections for: > > - Configuring Hardware Addresses (to account for new duplicate > hardware address handling). > - Observing IPMP interfaces with ipmpstat -i > - Machine-Parseable Output with ipmpstat -P > - IPv6 Link-Local Addresses (now split into two sections -- one > for data addresses, the other for test addresses). > > * Revised the definition of an "active" interface to no longer > consider duplicate hardware addresses (since in.mpathd now > offlines such interfaces automatically). > > [1] FWIW, all Nevada features with the exception of IPv6 appear to now > be operational.
Hi Meem, You (and anyone who assisted) did an incredible job with the IPMP design document. The document mentions that IPMP will be integrated into the fault management architecture at some point in time, and I was curious what type of integration is planned? Thanks again for the awesome document, - Ryan -- UNIX Administrator http://prefetch.net
