bstansberry: kkhan: a massive thing we are missing
[15:21] bstansberry: is ops on the domain root and server group
[15:21] bstansberry: start-servers, stop-servers, restart-servers
[15:22] bstansberry: so obvious I never even created a JIRA for it in 2010
[15:22] bstansberry: so of course I forget until the clustering guys asked about it!
[15:22] kkhan: bstansberry: ok, sounds good. I can add that
[15:22] kkhan: I'll create a JIRA
[15:23] bstansberry: I think the semantic for start is to start all servers that aren't started
[15:23] bstansberry: stop is stop all that are started
[15:23] bstansberry: kinda obvious
[15:23] kkhan: let me guess what restart does
[15:23] bstansberry: restart is to restart those that are already started
[15:24] bstansberry: i.e. it should not result in new servers running
[15:24] kkhan: Cool
[15:24] bstansberry: so the work is really all going to be in ServerOperationResolver
[15:24] kkhan: restart == reload or, stop + start?
[15:25] bstansberry: ah!
[15:25] bstansberry: excellent reminder
[15:25] bstansberry: maybe start with a full process restart
[15:25] bstansberry: but there's another JIRA to have a domain mode reload
[15:25] bstansberry: it doesn't work now
[15:25] kkhan: ok, I'll focus on restart now
[15:26] bstansberry: because a reload == a reparse and new set of ops to build
[15:26] bstansberry: nothing to parse in the domain
[15:26] darranl is now known as darranl_lunch.
[15:26] bstansberry: but if there was an op handler on the HC for a server reload, it potentially could send over the ops as part of a request to the server
[15:27] kkhan: ok, so how about I start on the new stuff and then take a look at the reload JIRA?
[15:27] bstansberry: yeah
[15:27] bstansberry: do the basics and then progress
[15:27] kkhan: BTW is this CR1 stuff, i.e. not needed today?
[15:27] bstansberry: absolutely