Erlang using up a lot of memory on idle node (a node where memcached was killed (oomkiller) but the node recovered)

In reference to : http://www.couchbase.com/issues/browse/MB-6962

The node: http://ec2-175-41-154-35.ap-southeast-1.compute.amazonaws.com:8091/
<Logs in S3, link in comments.>

erlang thread stats on the node:

(ns_1@ec2-175-41-154-35.ap-southeast-1.compute.amazonaws.com)1> erl_diag:print( [

{sort_by, total_heap_size}

, descending,

{limit, 10}

] ).

Process <4318.85.0>

initial_call :

{ale_disk_sink,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 347189360
reductions : 29775851

Process <4318.73.0>

initial_call : {ale_disk_sink,init,1}

status : waiting
message_queue_len : 0
total_heap_size : 222201240
reductions : 11145086

Process <4318.82.0>

initial_call :

{ale_disk_sink,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 91013672
reductions : 70029709

Process <4318.6.0>

initial_call : {gen_event,init_it,6}
status : waiting
message_queue_len : 0
total_heap_size : 37064392
reductions : 32649977

Process <4318.3143.0>

initial_call : {stats_reader,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 30408072
reductions : 48914847

Process <4318.18668.0>

initial_call : {stats_reader,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 16005168
reductions : 25817660

Process <4318.798.0>

initial_call : {ns_config,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 12341496
reductions : 98661442

Process <4318.12.0>

initial_call : {rpc,init,1}
status : waiting
message_queue_len : 0
total_heap_size : 10778048
reductions : 94581

Process <4318.94.0>

initial_call : {ale_disk_sink,init,1}

status : waiting
message_queue_len : 0
total_heap_size : 10770152
reductions : 30757005

Process <4318.796.0>

initial_call :

{gen_event,init_it,6}

status : waiting
message_queue_len : 0
total_heap_size : 10770152
reductions : 9544064