Liane Praza wrote: > Peter Memishian wrote: >> > i had a look at this over the weekend and have a prototype >> > fix for CR5087504 - it adds a 'refresh' subcommand to >> > svccfg, which, when an instance is selected, will take >> > a (new) running snapshot from the editing snapshot. the >> > error handling needs work, and i'm not sure that refresh >> > is the right name (since it may imply that the refresh >> > method is run also), but it may serve as a starting point. >> > i've tested it on an alternate repository (consisting of >> > the seed only) and a running repository, and it seems to >> > work fine. if i'm on the right track with this, i'd be happy >> > to push on with it if it'd help, but i'd need codereview/PSARC >> > fasttrack sponsorship help of course. >> >> Super, thanks for taking the initiative on this, Alan! >> >> Liane/David: could you offer some opinions on Alan's approach? >> >> One thing I'm unclear on is whether there would still be use-cases >> for a "refresh"-type subcommand once services are imported early in >> boot. If not, maybe the situation we're in is too fleeting to justify >> introducing a new subcommand? > > I'll take a look at the code in a bit, but yes -- I'm inclined to agree > that this should be a private subcommand for the timebeing with a quick > contract between our teams.
I went through the suggested fix in the bug report, and generally that was the approach I had been considering as well. In addition to making this private (we can quibble about naming later), I'd also do an explicit check to see if we're operating without SVCCFG_REPOSITORY set, and fail plus issue a message saying that svcadm refresh is the droid you're looking for. (I think that's the correct mechanism for avoiding the weird situation where the running service has been refreshed, but the restarter and the service itself hasn't yet been notified -- but I'm open to other suggestions too.) That can all wait, though, until it's been confirmed that this does actually solve both problems. Alan, you rock! liane