Results 337 comments of Keith Turner

> I think the design of this change is great so far but it is a major new feature, with a drastic increase in complexity, touching major parts of Accumulo...

@dlmarion made a few suggestions about the testing. I updated and the report and ran some new tests based on that. The new tests results are in the report.

> I'm thinking it would be better to leverage the scan hints to control a ScanServer-aware dispatcher, rather than add a new API for the consistency level. I think this...

> It's not the scan execution hints that are modifying the behavior... it's the configured dispatcher. And, the scan hints are still not affecting the data returned... it's the server...

> I'm also thinking that there should be a table configuration that enables/disables this feature. Currently, an admin can spin up some ScanServers and an application developer can enable eventual...

> Having this as behavior with an explicit API method to configure isn't any different. A configured dispatcher could just ignore that configuration and dispatch to an eventually consistent ScanServer...

> Is there a diagram or something to help illustrate the new dispatching design? Here is one. ![scan_server](https://user-images.githubusercontent.com/1268739/173731016-6c948612-52b7-4bba-9b08-ab0641f1a35a.png) The ScanServerDispatcher plugin is new and it runs client side to determine...

> The client side dispatcher concept is very different from the executor dispatching that is done in the tserver, but has a very similar name. It might be helpful to...

> It would be nice if the scan server dispatching inside the scan server worked the same as the existing per-table dispatcher inside the tserver. Ideally, the same dispatcher would...

It may be useful to add information to the following that informs a server side dispatcher if its running in a tserver or a scan server. Not sure about this....