Flavio Junqueira

I don't fully understand this issue. It refers to compatibility of the client API, but suggests to remove `pravegaVersion`. I'd like to understand what the precise problem is that we...

It's not clear to me that the suggested change is the best course of action. A stream job can be restarted, and if we use the same group name, then...

If we only have a single reader group through the lifetime of a Flink application, then we only really ever have one reader group to delete, so there is no...

I think I know what this issue is referring to, but it is missing specifics about what these end-to-end tests are.

> Pravega has a story to involve a Position for local recovery, but due to the async nature and segments may auto-scale, a group of positions in one reader group...

Thanks for pushing the limits of my memory, hopefully I'll be able to help out here. Local primary order is about the delivery order of messages. I had a quick...

@shiveshr what's the current story for internal streams and the API? I don't see any mention of it at least in the javadocs: https://github.com/pravega/pravega/blob/master/client/src/main/java/io/pravega/client/admin/StreamManager.java

Where in the code are we writing this message?

Three things: 1- @crazyzhou it would help if you explained in more detail and focused on the behavior you are trying to implement. My understanding is the following: upon recovery,...

@crazyzhou I still don't understand the behavior here: https://github.com/pravega/flink-connectors/pull/516/files#diff-077ca27a878ae584603740b5a82c4e4d826748be01c56649493eede64e3fb29dR87 With this code, we have: ``` try { readerGroup = readerGroupManager.getReaderGroup(readerGroupName); } catch (ReaderGroupNotFoundException e) { readerGroupManager.createReaderGroup(readerGroupName, readerGroupConfig); readerGroup = readerGroupManager.getReaderGroup(readerGroupName);...