Vincent Sevel
Vincent Sevel
any news on this?
any news on this?
Many years ago all of our EAPs were deployed as 2-nodes clusters, without a load balancer in front. Additionally, we also developed a proprietary protocol to allow ejb invocations over...
> Load balancer should automatically handle the connectivity errors based on the state of instances it is aware of. we have to recognize that awareness is a weak property here....
oh yes sure. I was certainly not saying that there is only one truth, and everything needed to be implemented to be correct. but knowing the use cases you want...
Part of the discussion is deciding if your load balancer is acting as an intermediate between the client and the service (that is the traditional definition), or a name resolver...
When we wrote our client side load balancing component, we also wrote a dedicated test suite to validate the behavior. We would run a client application (with the client side...
Interesting reading from the grpc project: [Transparent Retries](https://github.com/grpc/proposal/blob/master/A6-client-retries.md#transparent-retries) Grpc already distinguishes retries that it will do automatically (aka _transparent retries_) from retries that are governed by the retry policy. 3...
> Also, Stork has no idea if the operation you are going to call is idempotent or not, only the user and sometimes the transport layer know. so when the...
> At this point, I see @Retry as the correct way moving forward. if failover is addressed by MP FT, and not in stork or the rest client, then @Ladicek...