Here are my notes from the review, and of course these will prove to be much less comprehensive than the audio recording of the review itself. :-)
* Regarding gw-1, the project team will do an in-depth analysis of the potential use of SMF and reevaluate whether or not to store configuration in SMF. A detailed rationale for the resulting decision needs to be included in the case materials. The issues related to the UID used in exec_attr for ipadm, as well as the associated set of privileges (including file_dac_write) will be worked out with Gary Winiger. * Regarding jdc-1, the project team needs to clarify the design of the up-* ipadm subcommands and their relationship to the associated create-* subcommands. It was noted that there is a semantic difference between creating an object from scratch, and re-creating (or reloading, or instantiating, or whatever) an object from pre-existing persistent configuration (either at boot or when DR happens). Jim Carlson also questioned the need for multiple up-* subcommands, as well as their need to operate on individual objects. * Regarding jdc-4,7, ambiguity remains regarding the name of the object being manipulated by the subcommands that handle static addresses. Is it an address, a hostname, a label? During the meeting, Jim proposed that using labels for static addresses would remove any ambiguity between the object being manipulated by these subcommands. The project team noted that doing so would mean that something would need to keep track of the relationship between labels and the addresses currently configured on the system. The project team needs to investigate this and clarify. * Regarding jdc-5, the project team will do further investigation to determine if requiring CIDR notation for IPv4 as input for subcommands will be acceptable to administrators. * Regarding jdc-9, the project team will clarify this in documentation. * Regarding jdc-10, the function will be placed in libsocket. * Regarding jdc-11, the design of the IPMP-related ipadm subcommands needs to be worked out with the Clearview project prior to commitment. * Regarding jdc-16, the term "forwarding" will be used. -Seb
