Memcached connection errors even with --registry-exclude-image=*

[cmur2]

Describe the bug
I want to use the great fluxd only for syncing changes from git to a Kubernetes cluster but do not want to use the Docker image registry scanning feature of fluxd. Therefore I specified the --registry-exclude-image=* parameter which works.
To avoid having to run the now useless Memcached I removed it manually from the deployment descriptor yaml as this is another moving part that needs oversight and consumes resources.

Now I get error messages like although fluxd continues to work as usual:

ts=2019-11-06T08:38:37.409269301Z caller=memcached.go:156 component=memcached err="error updating memcache servers: lookup _memcached._tcp.memcached on 10.96.0.10:53: no such host"

To Reproduce
Steps to reproduce the behaviour:

1. My setup: fluxd on minikube with a Github repository and --registry-exclude-image=* parameter
2. Get deployment descriptor yamls from https://github.com/fluxcd/flux/tree/master/deploy excluding files starting with memcached-*
3. Modify https://github.com/fluxcd/flux/blob/master/deploy/flux-deployment.yaml to comment out all memcached-* options and include --registry-exclude-image=* parameter
4. Apply modified deployment descriptor yamls
5. See error in kubectl logs

Expected behavior
I would be great if fluxd would recognize the --registry-exclude-image=* parameter to stop trying to connect to Memcached without emitting errors.

Logs

ts=2019-11-06T08:38:37.409269301Z caller=memcached.go:156 component=memcached err="error updating memcache servers: lookup _memcached._tcp.memcached on 10.96.0.10:53: no such host"

Additional context
Add any other context about the problem here, e.g

• Flux version: 1.15.0
• Kubernetes version: 1.16 on minikube
• Git provider: Github
• Container registry provider: disabled

[texasbobs]

We are experiencing this as well. In the values.yaml file, we have set the memcache enabled to false

memcached:
  enabled: false

This correctly prevents the memcache files from being created. However, the output from helm has the parameter --memcached-hostname=name-flux-memcached listed for the flux deployment and Flux constantly produces the following message in the logs.

caller=memcached.go:156 component=memcached err="error updating memcache servers: lookup _memcached._tcp.name-flux-memcached on 169.254.25.10:53: no such host"

Even when removing the --memcached-hostname=, there is still a check for memcache servers and of course, they are not found.

We are one Flux version 1.16.0

Is there a way to prevent Flux from checking to see if a memcache server exists?

[masterkain]

│ ts=2019-12-08T08:27:50.821290777Z caller=memcached.go:112 component=memcached err="Fetching tag from memcache: read tcp 192.168.139.158:49960->10.100.168.47:11211: i/o timeout"

[2opremio]

This will be fixed by #2524

[texasbobs]

Thanks, @2opremio. I see it was last updated 23 days ago. What do you feel is the timeline for that PR?

[2opremio]

Christmas vacation got on its way. I will retake it next week.