Tailing oplog on secondary fails with CappedPositionLost

function repro() {

    uri="mongodb://localhost:27017/test?replicaSet=rs"

    # tail the oplog on the secondary

    mongo $uri --eval '

        var ol = db.getSiblingDB("local").oplog.rs

        var c = ol.find({}).readPref("secondaryPreferred").tailable({isAwaitData: true})

        var i = 0

        while (true) {

            c.forEach(function (o) {

                print(i++, o.wall)

            })

        }

    ' &

    sleep 5

    # generate some traffic

    mongo $uri --eval '

        while (true)

            db.c.insert({})

    '

}

2018-01-09T15:56:56.703-0500 E QUERY    [thread1] Error: getMore command failed: {

	"ok" : 0,

	"errmsg" : "PlanExecutor killed: CappedPositionLost: CollectionScan died due to failure to restore tailable cursor position. Last seen record id: RecordId(6509157867780571265)",

	"code" : 175,

	"codeName" : "QueryPlanKilled",

	"operationTime" : Timestamp(1515531416, 129),

	"$clusterTime" : {

		"clusterTime" : Timestamp(1515531416, 131),

		"signature" : {

			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),

			"keyId" : NumberLong(0)

		}

	}

} :