Karl Nilsson
Karl Nilsson
No this optimisation is for faster recovery when there are many Raft entries to be read from disk. Feel free to add a separate issue that describes your problem with...
@luos Is there anything concrete to address as part of this issue or can it be closed?
500ms for the aten poll_interval is low. I think it is 5s by default and in this case you probably want a larger value than the default That said that...
And yes the ra_log_meta data store _is_ slow. It uses `dets` internally. We just didn't have enough time to develop a decent alternative metadata storage engine. Perhaps that time has...
Hi yes I think it would be good to split this into multiple PRs: 1. for credit related changes 2. Message conversion related changes (ideally we want to consider `rabbit_msg_record`...
Hi yes I think it would be good to split this into multiple PRs: 1. for credit related changes 2. Message conversion related changes (ideally we want to consider `rabbit_msg_record`...
Ok so this is for the case where your destination address does not have a queue name segment, e.g. "/exchange/myexchange/" vs "/exchange/myexchange/myqueue". I think it is reasonable to make server...
my bad, you are quite right they are all server named queues here
If all your rabbit nodes aren’t clustered at the time of queue declaration they won’t be included in the quorum queue node selection. There is no way they could. Best...
Another alternative would be to simply catch any errors thrown by `apply/3` - at least this would be acceptable for pure state machines as there are no environmental factors to...