The genuine error aside, the io_lib part of that is because we don't
capture the stacktrace correctly when we log. I fixed this over on
cloudant a while back;

Couchdb does this;

        Tag:Error ->
            ?LOG_ERROR("Uncaught error in HTTP request: ~p",[{Tag, Error}]),
            ?LOG_INFO("Stacktrace: ~p",[erlang:get_stacktrace()]),
            send_error(HttpReq, Error)

The improvement is this;

        Tag:Error ->
            Stack = erlang:get_stacktrace(),
            twig:log(error, "req_err ~p:~p ~p", [Tag, Error,
                json_stack({Error, nil, Stack})]),
            send_error(HttpReq, {Error, nil, Stack})

This is fixed on master but not 1.1.x or 1.2.x. Worth a backport imo
if it's not too late for 1.2.0.

B.



On 19 March 2012 09:21, Dave Cottlehuber <dave@muse.net.nz> wrote:
> On 19 March 2012 09:08, Karoly Szanto <szanto.karoly@gmail.com> wrote:
>> Hi,
>> In the couchDB server log file I get the following weird error that I can't
>> understand
>>
>> [Thu, 08 Mar 2012 02:07:29 GMT] [error] [<0.1783.0>] Uncaught error in HTTP
>> request: {error,
>>                                                      {case_clause,
>>                                                       {error,enotsock}}}
>> [Thu, 08 Mar 2012 02:07:29 GMT] [info] [<0.1783.0>] Stacktrace:
>> [{io_lib_pretty,cind_tag_tuple,7},
>>                                 {io_lib_pretty,while_fail,3},
>>                                 {io_lib_pretty,print,6},
>>                                 {io_lib_format,build,3},
>>                                 {io_lib_format,build,3},
>>                                 {io_lib_format,build,3},
>>                                 {io_lib_format,build,3},
>>                                 {io_lib_format,build,3}]
>>
>> Any hints?
>>
>> Thank you,
>> Karoly
>
> Hi Karoly
>
> enotsock I think is complaining you passed a file not a socket.
> Nothing comes to mind immediately,
> but I'd be interested to know -
>
> what erlang version, OS are you running?
> do you have multiple network interfaces, and what is your bind_address
> set to for couchdb?
> any relevant couch activity prior to this issue? maybe some more log
> info would help.
> is this repeatable, and has couch been running correctly prior?
>
> Some reading -
> http://erlang.2086793.n4.nabble.com/enotsock-again-td2088701.html
> http://erlang.2086793.n4.nabble.com/erlang-questions-Unable-to-start-ejabberd-or-use-TCP-IP-communication-in-erlang-interpreter-td2093426.html
> http://comments.gmane.org/gmane.comp.lang.erlang.general/8410%3Cbr
>
> A+
> Dave