Gaël Bréard

Supposition (based on real observation) : - database load is high - On a persistent actor, 1 "persist" times out ... - .... but write order remains in queue (in...

Hum ok, in fact it's a well known subject : https://doc.akka.io/docs/akka/2.5/persistence-journals.html ``` * Please also note that requests for the highest sequence number may be made concurrently * to this...

Acknowledge mongo 3.0 replicaset = 1 primary and 3 secondary (1 is "hidden") Read preference is read on primary by default. it is overridden by the plug-in ?

Ok. But we still have a problem whatever the writeconcern if "persist" operation times out while the write on mongo is still on "queue" (not taken into account). And then...

I don't think we can retry persist: in doc of ```onPersistFailure``` we can read: ```The actor is always stopped after this method has been invoked.``` Thus retry would be to...

We don't have re-elections frequently, I was just trying to find a solution where overwriting could be acceptable (and thus in all cases - even with master changes - if...

We have our own implementation of split brain resolver. (implemented before OSS one is available) ...But the cluster itself is not split, and I've reproduced a complex situation where a...

Ok. Our main issue was a bad configuration of serializers that made journal event not replayed (only snapshots) for internal akka cluster sharding data and lead to complex situations... But...

This one is independent and may not be fixed bar #205 right ?

Hi @ivantopo and thanks for your reactivity ! We use openjdk 11 We saw the problem on centos-7. And on my laptop on ubuntu, each time my application starts I...