online data server crash: out of memory error

[sigstop]

With a sustained incoming data rate of about 20 interactions per second the data, the online analytic server crashes after about 3 days on a single machine with about 3 GB of RAM.

[sigstop@nci tapestry]$ erl -pa ebin
Erlang R16B02 (erts-5.10.3) [source-b44b726] [64-bit] [async-threads:10] [kernel-poll:false]

Eshell V5.10.3 (abort with ^G)
1> tapestry:start().

=INFO REPORT==== 1-Nov-2013::19:32:21 ===
Starting tapestry. View $TAPESTRY_HOME/log/console.log for operational messages.
<0.116.0>
tap_client_data: added client Pid = <0.1354.0>
2> loom_ofdp:get_all(default).
[<0.122.0>,<0.118.0>]
3> loom_ofdp:get_all(dns_tap).
[<0.123.0>,<0.119.0>]
tap_client_data: added client Pid = <0.29855.3>
tap_client_data: added client Pid = <0.7662.26>
tap_client_data: added client Pid = <0.4137.30>
tap_client_data: added client Pid = <0.3201.32>
tap_client_data: added client Pid = <0.4358.32>
tap_client_data: added client Pid = <0.15800.33>
tap_client_data: added client Pid = <0.1961.36>
tap_client_data: added client Pid = <0.30189.88>
tap_client_data: added client Pid = <0.17513.94>
4>
Crash dump was written to: erl_crash.dump
eheap_alloc: Cannot allocate 441485240 bytes of memory (of type "heap").
Aborted