> > Understood. Specific for this case, I will let Seb decide whether/how to > > provide the backwards bfu support for MAC-type plugin bits. > > I don't see how it will be possible to make backward bfu work in this > case unless we want to forever entrench the construct aggr0/<key> into > the configuration file syntax as opposed to the proposed aggr<key>. If > someone can come up with an idea, please let me know.
Usually, we've built this logic into BFU -- e.g., it could notice that a backward BFU is being done, and convert the file back to its old format. Later on, once we've crossed the "backward BFU grace period", the stuff can be hauled out of BFU. -- meem _______________________________________________ networking-discuss mailing list [email protected]
