[erlang-questions] My frustration with Erlang

Rapsey <>
Sat Sep 13 21:36:32 CEST 2008


How many mnesia transactions do you have left? From what I tested,
transactions are slow as hell. I'm sure you would get much better
performance, if a gen_server was in charge of a table and all queries were
dirty and went through that process. Unless you need transactions for more
than just thread safety. Then this method won't work of course.


Sergej

On Sat, Sep 13, 2008 at 8:59 PM, Joel Reymont <> wrote:

> I thought of running etop on another node to figure out what's going
> on at peak load. Sadly, etop crashes
>
>
> '1
> @mothership
> '                                                            17:19:03
>  Load:  cpu        90               Memory:  total     1309096
> binary       1302
>         procs   21259                        processes 1185824
> code         6406
>         runq        0                        atom          961
> ets         52340
>
> Pid            Name or Initial Func    Time    Reds  Memory    MsgQ
> Current Function
>
> ----------------------------------------------------------------------------------------
> <5624.3952.0>  proc_lib:init_p/5        '-' 1046280 1113656       0
> gen_server:loop/6
> <5624.0.0>     init                     '-'  576618   26320       0
> init:loop/1
> <5624.3941.0>  portServer3000           '-'  183416   88336       0
> tcp_server:socket_lo
> <5624.3956.0>  inet_gethost_native      '-'  178043    6864       0
> inet_gethost_native:
> <5624.73.0>    proc_lib:init_p/5        '-'   89051  688680       0
> disk_log:loop/1
> <5624.3937.0>  proc_lib:init_p/5        '-'   68600   24520       0
> gen_server:loop/6
> <5624.3938.0>  portServer6000           '-'   27716  418544       0
> tcp_server:socket_lo
> <5624.30.0>    group:server/3           '-'   20382   24520       0
> group:server_loop/3
> <5624.25.0>    file_server_2            '-'    7192   88400       0
> gen_server:loop/6
> <5624.16494.0> proc_lib:init_p/5        '-'    4770    8688       0
> gen_server:loop/6
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> =
> ========================================================================
>
> Output server crashed: connection_lost
>
> Another issue with etop:
>
> Erlang top got garbage {trace_ts,<5623.28.0>,out,
>                                  {user_drv,server_loop,5},
>                                  {1221,332097,210811}}
> Erlang top got garbage {trace_ts,<5623.25.0>,out,
>                                  {gen_server,loop,6},
>                                  {1221,332097,267093}}
> Erlang top got garbage {trace_ts,<5623.4.0>,out,
>                                  {gen_event,fetch_msg,5},
>                                  {1221,332097,340833}}
>
> I'm not doing any tracing myself so I really don't know where these
> traces come from. The easiest way to reproduce is to connect etop and
> then start Mnesia.
>
> --
> wagerlabs.com
>
> _______________________________________________
> erlang-questions mailing list
> 
> http://www.erlang.org/mailman/listinfo/erlang-questions
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20080913/61ec0320/attachment.html>


More information about the erlang-questions mailing list