ProcessOne Discussion Forums Copyright (c) 2013 ExpressionEngine tag:https:,2013:05:21 Web chat like Gmail or facebook - any setup recommendations? tag:https:,2010:/www.process-one.net/en/forum/viewthread/.396 2010-07-06T22:44:57Z jeans Hi,
we want to launch a web chat within a large social network with up to 50K users online simultaneously. Mostly like facebook, netlog or Gmail. Does anyone have recommendations? Are there any scalable plug and play solutions, including the JS client (no flash, no java, ...), where we just need to connect our users rosters?

Thanks in advance.

JD

]]> Install exmpp without sudo tag:https:,2011:/www.process-one.net/en/forum/viewthread/.609 2011-06-29T13:58:09Z Taddic Hi

I am new to this forum and I’ve just started doing my bachelor thesis at Uppsala University. My thesis is the creation of a serverless chat client (already been done yes, but it’s just a bachelor thesis, it’s not required it’s something new).

I would like to use exmpp, but the problem I’m having at the moment is that I want to use it on the Unix systems located at my school (Solaris machines) where I don’t have sudo rights. I can use make, autoreconf (as far as I could tell at least). I am not sure if they have expat installed or not, but I have asked them about this today so we’ll see about that. So is this possible or should I just give up using exmpp for this project (It’s not a requirement, I just wanted to use it to get some of the chat stuff for “free”, so I can focus on the “finding people” problem)?

]]> exmpp segfault tag:https:,2013:/www.process-one.net/en/forum/viewthread/.767 2013-03-27T18:42:00Z 2013-03-27T18:44:44Z argo Hello

in FreeBSD 9.1 amd64 with latest exmpp, erlang R15B03|R16B

I have a segfault

(gdb) backtrace
#0  bf_get_free_block (allctr=0x801873840, size=112, cand_blk=0x0, cand_size=0, flags=0) at beam/erl_bestfit_alloc.c:838
#1  0x0000000000440664 in mbc_alloc (allctr=0x801873840, size=Variable "size" is not available.) at beam/erl_alloc_util.c:1309
#2  0x0000000000440be0 in do_erts_alcu_alloc (type=174, extra=0x801873840, size=100) at beam/erl_alloc_util.c:3530
#3  0x0000000000440c57 in erts_alcu_alloc_thr_pref (type=174, extra=Variable "extra" is not available.) at beam/erl_alloc_util.c:3607
#4  0x0000000807169b83 in CRYPTO_malloc () from /lib/libcrypto.so.6
#5  0x00000008071457a4 in EVP_MD_CTX_copy_ex () from /lib/libcrypto.so.6
#6  0x0000000809108fef in tls1_final_finish_mac () from /usr/lib/libssl.so.6
#7  0x00000008091182e3 in ssl3_send_finished () from /usr/lib/libssl.so.6
#8  0x0000000809101f8c in ssl3_connect () from /usr/lib/libssl.so.6
#9  0x00000008090f3349 in ssl23_connect () from /usr/lib/libssl.so.6
#10 0x0000000808ed5782 in exmpp_tls_openssl_control (drv_data=0x80267d4a0, command=Variable "command" is not available.) at exmpp_tls_openssl.c:333
#11 0x000000000048f36d in erts_port_control (c_p=0x802782428, prt=0x80267d258, command=7, data=34442182898, retvalp=0x7fffff38cca8) at beam/io.c:3682
#12 0x0000000000503570 in erts_internal_port_control_3 (A__p=0x802782428, BIF__ARGS=0x802100040) at beam/erl_bif_port.c:250
#13 0x0000000000530d8c in process_main () at beam/beam_emu.c:2558
#14 0x00000000004a1c9c in sched_thread_func (vesdp=Variable "vesdp" is not available.) at beam/erl_process.c:5632
#15 0x00000000005b1c89 in thr_wrapper (vtwd=Variable "vtwd" is not available.) at pthread/ethread.c:106
#16 0x00000008010a20a4 in pthread_getprio () from /lib/libthr.so.3
#17 0x0000000000000000 in ?? ()
Cannot access memory at address 0x7fffff38d000 
]]> tweets dont deliver tag:https:,2013:/www.process-one.net/en/forum/viewthread/.762 2013-03-12T20:23:18Z petercoolkid my tweets dont deliver… have tried signing in again from tweet.im, buh i get d msg authentification failed

]]> [SOLVED] build exmpp failed on ubuntu 12.04 tag:https:,2013:/www.process-one.net/en/forum/viewthread/.758 2013-02-20T03:36:08Z 2013-02-22T07:53:39Z neokuno I want to give a try to ejabberd3, so, before that, I need install exmpp first, right?

First I download the latest stable version of exmpp, which is v0.9.9, from official github repo.

Followed step by step that in the README.md file:

cd exmpp-0.9.9
autoreconf -ivf
./configure 

Everything seems fine, but in the process of make, I got this error:

make  all-am
make[3]: Entering directory `/home/vagrant/exmpp-0.9.9/src'
{"init terminating in do_boot",{undef,[{make,all,[]},{erl_eval,do_apply,5},{erl_eval,expr,5},{init,start_it,1},{init,start_em,1}]}}

Crash dump was written to: erl_crash.dump
init terminating in do_boot ()
make[3]: *** [all-local] Error 1
make[3]: Leaving directory `/home/vagrant/exmpp-0.9.9/src'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/home/vagrant/exmpp-0.9.9/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/vagrant/exmpp-0.9.9'
make: *** [all] Error 2 

Could anyone help me figure this out?

Edit:
It seems I missing some dependencies that cause this problem. I dont know exactly what is missing, but after installed a lot of erlang package, eg. erlang-tools, erlang-docs…, I solved this issue.

]]> Avatar photo not showing up in pidgin? tag:https:,2013:/www.process-one.net/en/forum/viewthread/.753 2013-01-13T16:26:45Z slmingol I’m using ejabberd 2.1.5 and added a image to a user’s vcard/vcard2 like so:

% ejabberdctl set_vcard2 sam cat.server.me PHOTO EXTVAL 'http://www.gravatar.com/avatar/6a61a40ee46871d39e99835351e178ef?s=128&d=identicon&r=PG'
% ejabberdctl set_vcard sam chat.server.me PHOTO 'http://www.gravatar.com/avatar/6a61a40ee46871d39e99835351e178ef?s=128&d=identicon&r=PG' 

But my avatar never shows up in my pidgin client. Am I not understanding the abilities of vcard/vcard2 photos? Shouldn’t I be able to get them in this fashion from my xmpp client?

]]> Starting the graphical monitor from the ejabberdctl debug shell tag:https:,2012:/www.process-one.net/en/forum/viewthread/.750 2012-12-20T16:33:27Z 2012-12-20T19:21:49Z sjammula I am unable to get i:im() running from the ejabberdctl debug shell. That command works absolutely fine from a regular erlang shell. The Display environment variable is all set.
Is there a way that i can get this working? Is it even feasible?

]]> Ejabberd 2.1.11 User ungracefully exits and server fails to recognize user is offline for a long time tag:https:,2012:/www.process-one.net/en/forum/viewthread/.748 2012-12-14T18:19:49Z clander I am running ejabberd-2.1.11 and am experiencing something a bit weird. If a user has its internet connection dropped while it is connected and online I do not see the ejabberd server recognize this user to be offline for what looks like 2 hours. Is there something I can change configuration wise?

]]> Tsung XML tag:https:,2012:/www.process-one.net/en/forum/viewthread/.747 2012-12-11T06:54:44Z 2012-12-12T14:30:07Z OM I want to load test XMPP. I configure tsung xml
it authenticates and chats successfully.
but chat messages are randomly sent , all contacts from roster are not covered

How to configure tsung.xml to send chat messages to all roster contacts

I add users useruser1_1,useruser1_2,useruser1_3,useruser1_4,useruser1_5,
in each user’s roster all other contacts are added.

my tsung.xml

<?xml version=“1.0”?>
<!DOCTYPE tsung SYSTEM “/usr/share/tsung/tsung-1.0.dtd”>
<tsung loglevel=“info” version=“1.0”>

<clients>
      <client host=“cent6-64” use_controller_vm=“true”>
      <ip value=“192.168.0.53”>
      </client>
</clients>

<servers>
  <server host=“192.168.0.52” port=“5222” type=“tcp”></server>
</servers>

  <load duration=“3” unit=“minute”>
      <arrivalphase phase=“1” duration=“3” unit=“minute”>
          <users maxnumber=“5” arrivalrate=“1” unit=“second”></users>
      </arrivalphase>
      <user session=“xmpp-connection” start_time=“1” unit=“second”></user>
  </load>

<options>
  <option type=“ts_jabber” name=“global_number” value=“5”></option>
  <option type=“ts_jabber” name=“userid_max” value=“5”></option>
  <option type=“ts_jabber” name=“domain” value=“test.int”></option>
  <option type=“ts_jabber” name=“username” value=“useruser1_”></option>
  <option type=“ts_jabber” name=“passwd” value=“useruser1_”></option>
  <option type=“ts_jabber” name=“muc_service” value=“conference.test.int”>
</options>

<sessions>
  <session probability=“100” name=“xmpp-connection” type=“ts_jabber”>

  <request> <jabber type=“connect” ack=“no_ack”></request>

  <transaction name=“authenticate”>
    <request> <jabber type=“auth_get” ack=“local”></request>
    <request> <jabber type=“auth_set_plain” ack=“local”></request>
  </transaction>

  <request> <jabber type=“presence:initial” ack=“local”></request>

  <thinktime value=“5”>

  <for from=“1” to=“5” incr=“1” var=“counter”>

  <thinktime value=“1”>

        <transaction name=“online”>
          <request> <jabber type=“chat” ack=“no_ack” size=“10” destination=“online”></request>
        </transaction>

  <thinktime value=“1”>

        <transaction name=“online”>
          <request> <jabber type=“chat” ack=“no_ack” size=“10” destination=“online”></request>
        </transaction>

  <thinktime value=“1”>

        <transaction name=“online”>
          <request> <jabber type=“chat” ack=“no_ack” size=“10” destination=“online”></request>
        </transaction>

  <thinktime value=“1”>

        <transaction name=“online”>
          <request> <jabber type=“chat” ack=“no_ack” size=“10” destination=“online”></request>
        </transaction>

  <thinktime value=“1”>

        <transaction name=“online”>
          <request> <jabber type=“chat” ack=“no_ack” size=“10” destination=“online”></request>
        </transaction>
  </for>

  <thinktime value=“180”>

  <transaction name=“close”>
    <request> <jabber type=“close” ack=“no_ack”></request>
  </transaction>

  </session>

</sessions>
</tsung>

]]> MUC performance (delay) degradation tag:https:,2012:/www.process-one.net/en/forum/viewthread/.730 2012-09-26T17:06:20Z jlcruz Hello,

I’m using ejabberd to build an application to distribute messages in a one-to-many fashion where performance (delay) is critical. I opted for MUC vs PUBSUB because of its simplicity. One of the room participants sends a message so the rest of the room receives it.

I can understand that messages have to be delivered to all participants, therefore there will always be a delay between the first room participant being notified, and the last one.

However, I’ve found that delay increases linearly with the number of participants in the room. Even the delay for the first message leaving ejabberd to the first of the participants increases with the number of participants.

To be clear, I have measured the times between the XML message arrives to ejabberd to the moment the message is sent to the participants.
“(<0.427.0>:ejabberd_receiver:320) : Received XML on stream ...”
“(<0.428.0>:ejabberd_c2s:1553) : Send XML on stream ...”

Let me post some numbers to explain the case:


2 Participants Room
..............................First…......Last….......Diff
Average delay…...13.229…..14.197…..1.054


3 participants Room
..............................First…......Last….......Diff
Average delay…...15.940…..20.056…..4.020


4 participants Room
..............................First…......Last….......Diff
Average delay…...18.574…..26.377…...7.661


I’m using ejabberd-2.1.10, and my muc module config is:

  {mod_muc,      [{access, muc},
        {access_create, muc_admin},
        {access_admin, muc_admin},
              {history_size, 0},
        {max_users, 10000},
        {default_room_options, [
        {allow_change_subj, false},
        {allow_private_messages, false},
        {allow_query_users, false},
        {allow_user_invites, false},
        {members_by_default, false},
        {members_only, true},
        {moderated, false},
        {password_protected, false},
        {persistent, false},
        {public, false},
        {public_list, false}
          ]},
              {max_user_conferences, 1000000}]},

Is there any configuration change that can be done to improve this? When the number of participants in the room increases to 15, the delay goes up to 60ms, which is quite a lot.

Thanks in advance. Regards,
JL

]]>