Added data dir parameter while creating etcd container

[bhagyashree-sarawate]

Hi Sneha/William,
Added the ETCD_DATA_DIR parameter while creating the etcd container
Please review.
Verified whether the ETCD_DATA_DIR is set after running the playbook as follows:

[root@csimbe06-b13 ansible_3par_docker_plugin]# docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
6888fc2fa394 hpestorage/legacyvolumeplugin:3.2-prerelease "/bin/sh -c ./pluginâ¦" 6 minutes ago Up 6 minutes plugin_container
b04e082ca756 quay.io/coreos/etcd:v2.2.0 "/etcd" 7 minutes ago Up 7 minutes 2379-2380/tcp, 4001/tcp, 0.0.0.0:23790->23790/tcp, 0.0.0.0:23800->23800/tcp, 7001/tcp, 0.0.0.0:40010->40010/tcp etcd
809a2091a15a hello-world "/hello" 5 weeks ago Exited (0) 5 weeks ago epic_feynman

[root@csimbe06-b13 ansible_3par_docker_plugin]# docker exec -it b04e082ca756 /etcd
2019-06-10 09:25:01.639397 I | flags: recognized and used environment variable ETCD_ADVERTISE_CLIENT_URLS=http://10.50.0.160:23790,http://10.50.0.160:40010
2019-06-10 09:25:01.639466 I | flags: recognized and used environment variable ETCD_DATA_DIR=/var/lib/etcd
2019-06-10 09:25:01.639491 I | flags: recognized and used environment variable ETCD_INITIAL_ADVERTISE_PEER_URLS=http://10.50.0.160:23800
2019-06-10 09:25:01.639502 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER=etcd0=http://10.50.0.160:23800
2019-06-10 09:25:01.639539 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER_STATE=new
2019-06-10 09:25:01.639580 I | flags: recognized and used environment variable ETCD_INITIAL_CLUSTER_TOKEN=etcd-cluster-1
2019-06-10 09:25:01.639598 I | flags: recognized and used environment variable ETCD_LISTEN_CLIENT_URLS=http://0.0.0.0:23790,http://0.0.0.0:40010
2019-06-10 09:25:01.639610 I | flags: recognized and used environment variable ETCD_LISTEN_PEER_URLS=http://0.0.0.0:23800
2019-06-10 09:25:01.639632 I | flags: recognized and used environment variable ETCD_NAME=etcd0
2019-06-10 09:25:01.639724 I | etcdmain: etcd Version: 2.2.0
2019-06-10 09:25:01.639739 I | etcdmain: Git SHA: e4561dd
2019-06-10 09:25:01.639747 I | etcdmain: Go Version: go1.5
2019-06-10 09:25:01.639756 I | etcdmain: Go OS/Arch: linux/amd64
2019-06-10 09:25:01.639765 I | etcdmain: setting maximum number of CPUs to 32, total number of available CPUs is 32
2019-06-10 09:25:01.639850 N | etcdmain: the server is already initialized as member before, starting as etcd member...
2019-06-10 09:25:01.640081 C | etcdmain: listen tcp 0.0.0.0:23800: bind: address already in use

[wdurairaj]

Sorry. I noticed this late.

There should be a volume mapping like

docker run \
  -p 2379:2379 \
  -p 2380:2380 \
  --volume=${DATA_DIR}:/etcd-data \
  --name etcd ${REGISTRY}:latest \
  /usr/local/bin/etcd \
  --data-dir=/etcd-data --name node1 \

Your code change, only changes the --data-dir using the environment variable, meaning inside the etcd container the data of cluster is preserved.. but we need the backup on the etcd host, so a volume mapping will be required which is --volume=${DATA_DIR}:/etcd-data \

Request to add a volume mapping .

[wdurairaj]

Request to reference issue number in git commit message, so there will be an update on the corresponding bug as well

Eg.
git commit -m "Issue #611 -- added --data-dir as part of etcd container creation"

[wdurairaj]

Looks good for me. Let me see if Sneha has any comments.

[prablr79]

@sneharai4 can you quickly check and merge this change ?