juju.worker.peergrouper worker.go:211 cannot set replicaset: Found two member configurations with same _id field, members.0._id == members.1._id == 1

[Alvaro Uria]

* Juju 2.2.1, Xenial
* "juju enable-ha" has been run, but PRIMARY has no replicasets
* controllers #1 and #2 show as started and juju-db is up, however no machine-{1,2} users exist (so, unable to connect to slave mongos)
 - by stopping jujud-machine-{1,2}, stopping juju-db and (after backup) removing juju-mongodb3.2 deb pkg, such pkg is reinstalled after starting jujud-machine-{1,2} but status becomes the same (in "juju status --format yaml", they show up as "adding-vote"; while controller #0 "has-vote")
* machine-0.log shows the following error:
2017-07-13 07:17:10 ERROR juju.worker.peergrouper worker.go:211 cannot set replicaset: Found two member configurations with same _id field, members.0._id == members.1._id == 1

juju:PRIMARY> rs.status()
{
        "set" : "juju",
        "date" : ISODate("2017-07-13T07:17:59.088Z"),
        "myState" : 1,
        "term" : NumberLong(1),
        "heartbeatIntervalMillis" : NumberLong(2000),
        "members" : [
                {
                        "_id" : 1,
                        "name" : "127.0.0.1:37017",
                        "health" : 1,
                        "state" : 1,
                        "stateStr" : "PRIMARY",
                        "uptime" : 1721970,
                        "optime" : {
                                "ts" : Timestamp(1499930278, 44),
                                "t" : NumberLong(1)
                        },
                        "optimeDate" : ISODate("2017-07-13T07:17:58Z"),
                        "electionTime" : Timestamp(1498208312, 2),
                        "electionDate" : ISODate("2017-06-23T08:58:32Z"),
                        "configVersion" : 2,
                        "self" : true
                }
        ],
        "ok" : 1
}
juju:PRIMARY> rs.conf()
{
        "_id" : "juju",
        "version" : 2,
        "protocolVersion" : NumberLong(1),
        "members" : [
                {
                        "_id" : 1,
                        "host" : "127.0.0.1:37017",
                        "arbiterOnly" : false,
                        "buildIndexes" : true,
                        "hidden" : false,
                        "priority" : 1,
                        "tags" : {
                                "juju-machine-id" : "0"
                        },
                        "slaveDelay" : NumberLong(0),
                        "votes" : 1
                }
        ],
        "settings" : {
                "chainingAllowed" : true,
                "heartbeatIntervalMillis" : 2000,
                "heartbeatTimeoutSecs" : 10,
                "electionTimeoutMillis" : 10000,
                "getLastErrorModes" : {

                },
                "getLastErrorDefaults" : {
                        "w" : 1,
                        "wtimeout" : 0
                },
                "replicaSetId" : ObjectId("594cd837331fb00c0fa13128")
        }
}

On a test environment, I checked that "juju enable-ha" creates mongo replicasets with ids 1, 2 and 3 (for controllers #0, #1 and #2, respectively).

Please let me know if you need further information.

[Alvaro Uria (aluria)]

Such test env uses an openstack endpoint (keystone auth, nova instances).

Issue related to bug 1704072