amp cluster create command starts too many services

[subfuzion]

These services shouldn't be started (and no reason for 3 replicas of pinger).

$ amp cluster create
[user tony @ localhost:50101]
. . .
ID                  NAME                MODE                REPLICAS            IMAGE                                    PORTS
ss4cxe8q7i31        vis_portainer       replicated          1/1                 portainer/portainer:latest
wi41ujisapav        vis_visualizer      replicated          1/1                 manomarks/visualizer:latest
zszm5dji5ywc        pinger_pinger       replicated          3/3                 subfuzion/pinger:latest

[subfuzion]

Note also that this fails in any case because visualizer has moved:

wi41ujisapav        vis_visualizer      replicated          0/1                 manomarks/visualizer:latest

The new repo is: dockersamples/visualizer

[subfuzion]

@ndegory I was under the impression you handle last week for the release. Are you still going to take care of it or do you want me to? You moved the stack itself from platform/stacks to examples, but this isn't reflected yet with the bootstrap image and right now cluster create fails (due to visualizer).

[ndegory]

@subfuzion the amp cluster create doesn't deploy the examples.
Here's the list I got after a fresh deployment:

ID                  NAME                   MODE                REPLICAS            IMAGE                                          PORTS
1hdxne9me94f        amp_nats               replicated          1/1                 appcelerator/amp-nats-streaming:v0.5.0
5ghiteb9m6zv        amp_etcd               replicated          1/1                 appcelerator/etcd:3.1.9
661qyut2ztat        amp_gateway            global              1/1                 appcelerator/gateway:0.13.0-dev
9o6avb6nset6        amp_node_exporter      global              1/1                 prom/node-exporter:v0.14.0                     *:9100->9100/tcp
e3ezjo7zkx9q        amp_amplifier          global              1/1                 appcelerator/amplifier:0.13.0-dev              *:50101->50101/tcp
iyjd9lukm9nn        amp_elasticsearch      replicated          1/1                 appcelerator/elasticsearch-amp:5.4.2
lmq4wbz0h06f        amp_kibana             replicated          1/1                 appcelerator/kibana:5.4.2
lnjqsz3sc24v        amp_nats_exporter      replicated          1/1                 appcelerator/prometheus-nats-exporter:latest
lpwj3u8rzwm2        amp_prometheus         replicated          1/1                 appcelerator/amp-prometheus:0.13.0-dev         *:9090->9090/tcp
nqv77y40wbz3        amp_ampbeat            replicated          1/1                 appcelerator/ampbeat:0.13.0-dev
nv82mzcof09m        amp_alertmanager       replicated          1/1                 prom/alertmanager:v0.7.1                       *:9093->9093/tcp
okfuo1n2a7f2        amp_grafana            replicated          1/1                 appcelerator/grafana-amp:1.2.3
p8dbyxgmavqe        amp_haproxy_exporter   replicated          1/1                 prom/haproxy-exporter:v0.7.1
q6fplc7h8chv        amp_agent              global              1/1                 appcelerator/agent:0.13.0-dev
vhvwejm4udf5        amp_portal             replicated          1/1                 appcelerator/portal:latest
y5wnhm4uof1r        amp_proxy              global              1/1                 dockercloud/haproxy                            *:80->80/tcp,*:443->443/tcp,*:1936->1936/tcp

I guess the example services you see are from an older deployment.
When you destroy your local cluster (sic), it only removes the core services, not the user stacks you may have deployed.

[subfuzion]

@ndegory I have no deployed services / stacks. I ensured nothing was running, then ran docker system prune, then did a fresh build, then ran the command. However, I'm removing all images now to ensure only those that are needed are pulled (and up-to-date).

[subfuzion]

I understand the issue now and will open a new issue that is more appropriate for preventing this behavior going forward. I will put the details in the new issue and reference this one.