These are chat archives for akkadotnet/akka.net

10th
Mar 2018
Onur Gumus
@OnurGumus
Mar 10 2018 01:51
Should I be using at least once delivery with cluster sharding to where fault tolerance?
Bartosz Sypytkowski
@Horusiath
Mar 10 2018 07:34
@mtmk you'd need to track that by yourself. I remember I did a pattern like the one you've looking for in the past, but I don't think I'll be able to find it right now (it was around 3 years ago).
but in general you'd need a custom parent actor, that works like a reaper and gateway in one (it kills most idle children and all messages to those children should go through it, so he can determine which ones are the most idle)
@Efp95 never had such a problem. How are you setting up the fallbacks? They always should start from most detailed to most abstract ones (so plugin configs always work as the fallbacks of your own config)
Bartosz Sypytkowski
@Horusiath
Mar 10 2018 07:39
@OnurGumus there's nothing wrong with that. Just keep in mind, for the entities, actor path can change over time. To make it work, the path of your delivery target should be sharding region, while the message is a shard envelope with a message to a given entity.
Onur Gumus
@OnurGumus
Mar 10 2018 07:40
Right
@Horusiath have you used things like ncache?
Onur Gumus
@OnurGumus
Mar 10 2018 09:02
Are remotely deployed actors still supervised by their remote parents?