: when reloading a core, it seems that the execution of firstSearcher and : newSearcher events will happen after the new core takes over from the : old. This will effectively stall querying until the caches on the new : core are warmed (which can take quite a long time on large : installations). Is this correctly observed, and if so, is there a way to : postpone the core swap?
core reloading ensures that the old core isn't closed until it's finished handling old requetss, but new requests are sent to the new core as soon as it starts up. It's effectively the same as an UNLOAD and a CREATE except there is no window of time where *no* core is available using that name. If you want there to be overlap, and do warming on the new core before it gets any traffic, you should consider explicitly using CREATE + SWAP + UNLOAD which gives you total control over when the SWAP happens. -Hoss
