Matthieu Huguet

> I have no experience with ES clusters, but what would be the expected behavior if we upgrade the ES release with a rolling update, _i.e._ a cluster (pods) with...

I reopen this issue because the last part is not done yet : > After this is done, modify existing apps to make use of this HA Elasticsearch

I've just got the same issue with a non blue/green application. My DeploymentConfig definition was something like this : ``` apiVersion: v1 kind: DeploymentConfig metadata: labels: app: elasticsearch service: elasticsearch...

> What if we use the `force: true` flag during DC creation/patching? It doesn't work (see @lunika's comment in *Possible solution*) I also tried `merge_type: merge`, `merge_type: strategic-merge` and `apply:...

If we do it that way (delete / create), all interesting features of the DeploymentConfig (like rollback or deployment strategies) will not be usable. What do you think about the...

You're right, we lose the version label in pods but we keep it in deploymentconfig.

Idea : try to add `-w` option to nc command in the sidecar container.

Maybe we can use kubernetes to store the deployment state of our applications, instead of deploying a new app only for this purpose ? K8s uses etcd internally as a...


Hi @rexon07, Maybe the machine on which you ran the `bbb-stress-test` script is overloaded? It's quite resource-intensive and I used to run it on multiple VMs.