nickva commented on PR #5807: URL: https://github.com/apache/couchdb/pull/5807#issuecomment-3612772385
> A cluster should tolerate a minority of nodes in mm without this leaking into replies (i.e N nodes in a 2N+1 node cluster, and specifically those N nodes hosting any of the attempted doc). That's what what the PR does -- it hides the mm mode unless there is only mm mode left as the error. What should we emit to the user if all nodes are in mm? It seems odd to invent yet another type of error, an indication that all the shard copies are on nodes in mm seems most helpful in diagnosing the issue. > singling out maintenance_mode in particular feels wrong (though we certainly don't want to expose it in responses). We single it out for all the streaming requests - views, changes, etc, because that's when we start replacements. https://github.com/apache/couchdb/blob/fe3fc9fffd3a79a568dfaf99676c6d0ebeb291e1/src/fabric/src/fabric_streams.erl#L134-L137 If we didn't have a common fabric_streams abstraction we'd have 4 or 5 clauses to specially handle it in the code, in other words it's definitely a bit of a "special" case. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
