HTTP polling is working. Clients can logon/off and message via http polling. However, we are having frequent ejabberd failures created by users connecting via http polling. Cannot quite decipher the error log results. The failures put entries in the sasl.log and ejabberd.log. Failure excerpts are below. Thanks in advance for any help. Anyone have pointers to the root cause?
Environment: ejabberd 1.1.3 on CentOS.
SASL Log Output
=CRASH REPORT==== 7-Apr-2007::09:56:23 ===
crasher:
pid: <0.1571.0>
registered_name: []
error_info: {function_clause,[{xml,
get_tag_attr,
["from",{error,parse_error}]},
{ejabberd_http_poll,get_jid,2},
{ejabberd_http_poll,resend_message,1},
{lists,foreach,2},
{ejabberd_http_poll,terminate,3},
{gen_fsm,terminate,7},
{proc_lib,init_p,5}]}
initial_call: {gen,init_it,
[gen_fsm,
<0.208.0>,
<0.208.0>,
ejabberd_http_poll,
["15446df2ed7a8193bbd427a185f1a914ce480397",[]],
[]]}
ancestors: [ejabberd_http_poll_sup,ejabberd_sup,<0.36.0>]
messages: []
links: [<0.65.0>,<0.208.0>]
dictionary: []
trap_exit: false
status: running
heap_size: 2584
stack_size: 21
reductions: 31077
neighbours:
ejabberd.log output
=ERROR REPORT==== 2007-04-09 17:40:08 ===
** State machine <0.465.0> terminating
** Last message in was {timeout,#Ref<0.0.0.5697>,[]}
** When State == loop
** Data == {state,"63f9553fcc7dc0c814f245e8841cf43c47426fe3",
"W/mhjQzcX6Hr1A0vzVxRNkuD0+U=",
["Replaced by new connection"],
[],
{<0.467.0>,#Ref<0.0.0.5698>},
<0.467.0>,
#Ref<0.0.0.5697>}
** Reason for termination =
** {function_clause,[{xml,get_tag_attr,["from",{error,parse_error}]},
{ejabberd_http_poll,get_jid,2},
{ejabberd_http_poll,resend_message,1},
{lists,foreach,2},
{ejabberd_http_poll,terminate,3},
{gen_fsm,terminate,7},
{proc_lib,init_p,5}]}
