sample: update local-drone-control-scala for kubernetes #1015
Conversation
| @@ -8,6 +8,7 @@ include "cluster" | |||
|
|
|||
| akka { | |||
| loglevel = DEBUG | |||
| remote.artery.canonical.hostname = "127.0.0.1" | |||
There was a problem hiding this comment.
do we have to define this at all? doesn't work with default?
There was a problem hiding this comment.
Will try it, was just retaining the config that was there before.
There was a problem hiding this comment.
Yeah, shouldn't be needed. And single-node cluster.
| @@ -0,0 +1,39 @@ | |||
| # Production configuration for running the local-drone-control service in Kubernetes, | |||
There was a problem hiding this comment.
Maybe we should not name this prod.conf since the single node application.conf is also a prod alternative. Maybe application-kubernetes.conf?
There was a problem hiding this comment.
Sounds good. Was going to name it kubernetes.conf originally.
There was a problem hiding this comment.
Although, I guess the single node with H2 version can run in kubernetes too.
There was a problem hiding this comment.
true, but then we should have rather different deployment yml for that so that it's not forming a cluster with other nodes
| canonical.port = 2552 | ||
| canonical.port = ${?REMOTE_PORT} | ||
| } | ||
|
|
There was a problem hiding this comment.
How does this config file compare to other clustered samples? e.g. restaurant-drone-deliveries-service-scala
I'm missing for example downing-provider-class
There was a problem hiding this comment.
Ok. I'll make sure it has everything. And it could make sense to consolidate into the cluster.conf file for when it's clustered, (and not included from single-node application.conf).
| @@ -1,5 +1,3 @@ | |||
| akka { | |||
| actor.provider = cluster | |||
There was a problem hiding this comment.
maybe this config file is just confusing for this sample, and we can put this in application.conf and application-kubernetes.conf ?
There was a problem hiding this comment.
We could put more cluster config here, for the clustered variants, and then only have provider = cluster in the default single-node application.conf.
There was a problem hiding this comment.
ok, I think something is missing for the single node one as well. contact-point-discovery config should be there. That is now only in local-shared.conf. The local files are intended for local dev mode.
There was a problem hiding this comment.
Single-node has its own main, does a self-join there, and doesn't start cluster bootstrap. Local has config discovery.
There was a problem hiding this comment.
alright, I'll look over the stuff. confusing to use different approaches, but nothing you should spend time on in this PR
|
Config reworked for suggestions. Native image config updated for changes (SBR included). Tested in k3s again. |
Update local-drone-control config to run in kubernetes. Run with the native image tracing agent to generate updated native image config. Tested with k3s.