No Voice participants show up

83 views
Skip to first unread message

gethanos

unread,
Nov 22, 2009, 6:44:12 PM11/22/09
to BigBlueButton-dev
hi.
I tried to figure out what the problem was, but still no luck.I am
running Ubuntu 9.04 with bbb 0.62.So here it is.

1.When someone click on Voice enable button, he gets the notification
and he can talk/listen to others.
2.Problem is that the Voice participants List is empty. (Works only
the first time, after a clean installation)

I am working on Parallels Desktop (VM), so it was easy to make clean
installations of Ubuntu with bbb to test.The first time that i am
logging in, the Voice participants list shows up correctly.After that,
i am not able to see anyone in that list, and tried from many
different PCs.

Thanks for your help and time.

gethanos

unread,
Nov 22, 2009, 7:00:16 PM11/22/09
to BigBlueButton-dev
I don't know if that helps, but I am getting those lines inside "/usr/
share/red5/log/error.log"

"2009-11-23 01:57:26,498 [NioProcessor-1] WARN
org.red5.server.Context - Bean lookup failed for
participantsSO..soservice in the application
context"

Richard Alam

unread,
Nov 22, 2009, 7:19:20 PM11/22/09
to bigblueb...@googlegroups.com
can you please post what's in the /usr/share/red5/log/bigbluebutton.log?
> --
>
> You received this message because you are subscribed to the Google Groups "BigBlueButton-dev" group.
> To post to this group, send email to bigblueb...@googlegroups.com.
> To unsubscribe from this group, send email to bigbluebutton-...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/bigbluebutton-dev?hl=.
>
>
>



--
---
BigBlueButton
http://www.bigbluebutton.org
http://code.google.com/p/bigbluebutton

gethanos

unread,
Nov 22, 2009, 7:28:05 PM11/22/09
to BigBlueButton-dev
Thanks Richard for your reply.
Well, i found my answer in that post
http://groups.google.com/group/bigbluebutton-dev/browse_thread/thread/9dcec995f0e98259#

It seems that there is a problem with services start on reboot, so it
works after you restart them manually.
If you still need the "bigbluebutton.log" i can attach it.
Is there any solution about the "services issue"?

Thanks again.

Tamás

unread,
Nov 22, 2009, 8:56:16 PM11/22/09
to BigBlueButton-dev
Hi Richard,

I have the same bean lookup failure now. I added a comment to issue
#237 where I complained about the voice part does not work. Now
somehow it works again (I just did the same as earlier... restarted
all services), there is no functional problem currently, but I see in
the error.log the followings:

2009-11-23 03:12:48,972 [NioProcessor-2] WARN org.red5.server.Context
- Bean lookup failed for participantsSO..soservice in the application
context
2009-11-23 03:12:53,977 [Asterisk-Java DaemonPool-1-thread-1] WARN
org.red5.server.Context - Bean lookup failed for
meetMeUsersSO..soservice in the application context

The contents of my error.log, sip.log, red5.log and bigbluebutton.log
can be found here:
http://pastebin.com/f3a37ff8a

Could you please help what is the problem? And please try to answer to
my earlier question as well: http://code.google.com/p/bigbluebutton/issues/detail?id=237

Thank you very much. BTW this project is one of the best open source
projects I've seen (well documented, user- and developerfriendly,
clear source code, separated modules, good support), thank you for
your great work.

Best regards,
Tamás Fitos

Walter Tak

unread,
Nov 23, 2009, 1:34:06 AM11/23/09
to bigblueb...@googlegroups.com
Well after , couch , a day of puzzling with vmware's vmplayer that didn't
want to work in with vncserver and then fiddling with the custom
installation-instructions found on
http://code.google.com/p/bigbluebutton/wiki/InstallingBigBlueButtonCentOS ,
having to compile dahdi and asterisk myself because the basic packages of
Debian just didn't want to work , ending up reading
http://code.google.com/p/bigbluebutton/wiki/InstallingBigBlueButton to see
if I missed something, redoing several parts I *finally got it to work*.
Hooray.

Obviously I overcame all hurdles ; Ubuntu has different package-names than
Debian. Debian has different versions marked as 'stable' (and other marked
as unstable or testing that were broken and couldn't be installed). Besides
that I had to recompile dahdi and asterisk several times because a module
"app_meetme" didn't want to compile. (Asterisk said it couldn't load the
Meetme app). The config of Asterisk showed that the meetme app had failed
dependencies. Arf.

After looking in the source (of Asterisk) it seemed I was missing
dahdi-source code (for which I installed a Debian package, BEEEP, not good
enough ; had to install it from the source, compile it and THEN compile
asterisk.

After that Asterisk refused to "work" in the conferencing system ; All I
heard was "invalid conference blabla" (voice). In the logs
(/var/log/asterisk/messages) I found lines like:

[Nov 23 07:04:10] WARNING[4302] app_meetme.c: Unable to open pseudo device

After many searches on the list and endless recompiles of dahdi I found out
that I had to actually execute a command "dahdi start" which inserts the
freshly compiled kernel modules. However that command is nowhere in the
docs. Ahhh. After a restart of Asterisk no more "pseudo device" errors
showed up in the Asterix logfiles.

Then reconnected to the app and yes, when pressing the "headphones" icon I
heard some chick talk sweet in my ear, playing some music as if I were in
some elevator and when I opened a second browser I was in the voice
conference room. Woot.

Nice stuff Fred and Richard. Kinda complex to setup though even for an old
fart like me that used to compile Slackware 1.2.13 kernels back in the days
;)

(The problem with the VMWare thing is that I do have a dedicated Linux
server for testing but I don't have xwindows installed on the box nor a
monitor/display, just vnc4server for remote control, if SSH would fail.

When I tried to install VMWare Player that piece of software kept crashing
when starting the VM you made ; it's a VMplayer problem ; "xrandr error on
display blabla." ; the used free vnc4server didn't support xrandr (for
dynamic resizing of a desktop) and that was the end of the "easy VM tour".

I do experience the small bug that no voice participants are being shown ;
Red5 shows this error (which others reported as well) :

2009-11-23 07:08:00,376 [NioProcessor-1] WARN org.red5.server.Context -
Bean lookup failed for participantsSO..soservice in the application context

Using the "recommended" 0.8RC3 version btw no trunk or 0.9 variants.

Walter

Walter Tak

unread,
Nov 23, 2009, 1:57:03 AM11/23/09
to bigblueb...@googlegroups.com
Oh btw I fixed that Red5 error "Bean lookup failed for
participantsSO..soservice in the application" by installing the 0.8.0 final
version of Red5 instead of the 0.8RC2 version.

0.9RC2 didn't work btw ; lots of errors on startup ; looks like the webapps
(esp. sip) aren't compatible, might be due to the dreaded loggers.

With 0.8.0 you do see other voice participants (aka users). Works like a
charm right now.

Regards,
Walter
Tam�s Fitos

gethanos

unread,
Nov 23, 2009, 5:46:43 AM11/23/09
to BigBlueButton-dev
Thanks alot Walter :)
I will try it soon :)

Hugo Flambó

unread,
Nov 23, 2009, 6:55:41 AM11/23/09
to bigblueb...@googlegroups.com
Hi Walter,

I've been faced with this exactly problem on Amazon EC2

Dahdi started and Asterisk running... but on logs always see that error:
app_meetme.c: Unable to open pseudo device

I remember that I checked that device exists, daemons have
permissions to open the device, dahdi was working as expected,
asterisk working ...all working but always got that error when I tried
the bbb voice app and ended to let it parked till I have some time
again to try it again....

What do you mean with

" I found out that I had to actually execute a command "dahdi start"
which inserts the freshly compiled kernel modules. However that
command is nowhere in the docs"

That command is not the one to start dahdi? i guess I've executed it a
couple of time... dahdi stop , dahdi start ... I've check that it was
running and that dahdi_dummy was loaded...


Can you kind give some hints to me on that?

Those where some of my tests:

Asterisk console I got:

Verbosity is at least 5
-- Registered SIP '3011' at 127.0.0.1 port 5081 expires 3600
-- Saved useragent "mjsip stack 1.6" for peer 3011
[Nov 11 12:20:07] NOTICE[4778]: chan_sip.c:12669
handle_response_peerpoke: Peer '3011' is now Reachable. (1ms / 2000ms)
-- Executing [85115@bbb-voip:1] MeetMe("SIP/3011-081f95f8",
"85115|cdMsT") in new stack
[Nov 11 12:20:07] WARNING[11679]: app_meetme.c:774 build_conf: Unable
to open pseudo device
-- <SIP/3011-081f95f8> Playing 'conf-invalid' (language 'en')
== Spawn extension (bbb-voip, 85115, 1) exited non-zero on 'SIP/3011-081f95f8'


I guess the problem is app_meetme.c:774 build_conf: Unable to open
pseudo device , not sure what is about but at first guess I with
dahdi_dummy wich is currently loaded and running...

executing: modinfo dahdi_dummy

I GET:

filename: /lib/modules/2.6.18-xenU-ec2-v1.0/dahdi/dahdi_dummy.ko
license: GPL v2
author: Robert Pleh <rober...@hermes.si>
description: Timing-Only Driver
depends: dahdi
vermagic: 2.6.18-xenU-ec2-v1.0 SMP mod_unload 686 REGPARM gcc-4.0
parm: debug:int

executing: dahdi_test -v
Opened pseudo dahdi interface, measuring accuracy...

8192 samples in 8190.560 system clock sample intervals (99.982%)
8192 samples in 8190.096 system clock sample intervals (99.977%)
8192 samples in 8182.424 system clock sample intervals (99.883%)
8192 samples in 8182.376 system clock sample intervals (99.883%)
8192 samples in 8222.616 system clock sample intervals (100.374%)
8192 samples in 8174.175 system clock sample intervals (99.782%)
8192 samples in 8190.512 system clock sample intervals (99.982%)
8192 samples in 8198.399 system clock sample intervals (100.078%)
8192 samples in 8182.384 system clock sample intervals (99.883%)
8192 samples in 8190.480 system clock sample intervals (99.981%)
8192 samples in 8190.440 system clock sample intervals (99.981%)
--- Results after 11 passes ---
Best: 99.982 -- Worst: 99.626 -- Average: 99.898383, Difference: 99.980538

executing: ls /sys/module/
capability commoncap crc_ccitt dahdi dahdi_dummy i8042 keyboard
lockd md_mod mousedev nfs printk tcp_bic

executing: ls /dev/dahdi/
channel ctl pseudo timer


executing: dahdi_cfg -v
DAHDI Tools Version - 2.2.0

DAHDI Version: 2.2.0.2
Echo Canceller(s):
Configuration
======================


0 channels to configure.


So dahdi_tools must be working too..


Regards,

Hugo

Walter Tak

unread,
Nov 23, 2009, 10:11:53 AM11/23/09
to bigblueb...@googlegroups.com
hey Hugo,

when I compiled dahdi from source that process created init scripts in my
/etc/init.d/ directory (Debian) ; with "/etc/init.d/dahdi start" you
normally start a Debian service.

That script inserted/loaded the dahdi compiled kernel modules.

Some snippets from the startup script [/etc/init.d/dahdi] :

...
echo "Loading DAHDI hardware modules:"
modprobe dahdi

...

if [ ! -e /proc/dahdi/1 ]; then
echo "No hardware timing source found in /proc/dahdi,
loading dahdi_dummy"
modprobe dahdi_dummy 2> /dev/null
fi

...

You can check if module "dahdi" really is running with "lsmod" ; my output:

*****
dahdi_dummy 3204 0
dahdi_transcode 6408 1 wctc4xxp
dahdi_voicebus 32000 2 wctdm24xxp,wcte12xp
dahdi 184552 14
dahdi_dummy,xpp,dahdi_transcode,wcb4xxp,wctdm,wcfxo,wctdm24xxp,wcte11xp,wct1xxp,wcte12xp,dahdi_voicebus,wct4xxp
crc_ccitt 2080 1 dahdi

*****

This shows that "dahdi_dummy" is loaded but not in use ( 0 / zero ) but the
other devices/modules like dahdi_transcode , which is required , is in use
as well as module "dahdi".


Hope this helps,

Walter

PS remember I have less than 24 hours of experience with dahdi, asterisk and
BBB so don't expect any expert answers yet ;)

Richard Alam

unread,
Nov 23, 2009, 10:28:30 AM11/23/09
to bigblueb...@googlegroups.com
Great job Walter. Glad you got it going :)

I thought when you do a "make config" for dahdi, it sets up the
scripts to start autmatically. But perhaps, you need to run it
manually the first time. The scripts sets up auto start after reboots.

Hopefully, we'll be able to remove the dahdi dependency this week.
That's what I'll be working on...moving to app_konference instead of
app_meetme.

I thought we had the 0.8final release. I'll double check and upgrade
to the final one. But probably not 0.9 yet as I haven't looked what
needs to be changed to move to that version.

Richard

Richard Alam

unread,
Nov 23, 2009, 10:43:59 AM11/23/09
to bigblueb...@googlegroups.com
Hi,

Please let us know if this fixes your problem or if it still persists.

I'll be looking more into the problem once I refactor the voice
application to support app-konference.

Richard

Hugo Flambó

unread,
Nov 23, 2009, 11:00:50 AM11/23/09
to bigblueb...@googlegroups.com
Hey Walter...

hummmm yes I've checked all and I'm sure it was running .... not sure
about of that 0 flag for in use or not...but will check when I have
time

you say:
"This shows that "dahdi_dummy" is loaded but not in use ( 0 / zero ) but the
> other devices/modules like dahdi_transcode , which is required , is in use
> as well as module "dahdi"."

dahdi_transcode is required? I think I've done something like this:

mv /etc/dahdi/modules /etc/dahdi/modules.backup
echo "# Empty conf file in order to load only dahdi_dummy" > /etc/dahdi/modules

Are you sure that transcoder module is required? I've checked dahdi
documentation and wctc4xxp that is a digital card
Since I was running on Virtualized env I think only dummy module is
required.. (since there's no hw)

Can you test it without any other module loaded and check if it still
works (procedure above) ...?? If not that's ok...but I recommend it to
you if you don't have telephony hw ;)

thanks anyway

BR

Hugo

1.1. Digital Cards
wct4xxp:
Digium TE205P/TE207P/TE210P/TE212P: PCI dual-port T1/E1/J1
Digium TE405P/TE407P/TE410P/TE412P: PCI quad-port T1/E1/J1
Digium TE220: PCI-Express dual-port T1/E1/J1
Digium TE420: PCI-Express quad-port T1/E1/J1

Richard Alam

unread,
Nov 23, 2009, 11:09:31 AM11/23/09
to bigblueb...@googlegroups.com
Hi Hugo,

Yes, you should be only using dahdi_dummy.

Richard

Hugo Flambó

unread,
Nov 23, 2009, 11:18:54 AM11/23/09
to bigblueb...@googlegroups.com
Thanks Richard,

I was 99% sure that I was right....
One of this days I'll try it again... not for now...

I've seen some post related to flex ui with release 0.62

just to say that when release 0.62 have come out... I've compiled ui
too and it was displaying 0.4...

I've forget to ask <about it.. :S

BR Hugo

Richard Alam

unread,
Nov 23, 2009, 11:22:57 AM11/23/09
to bigblueb...@googlegroups.com
When you compile your client, it's going to display 0.4 (we should
have changed it:)).

You get a "date formatted" version when you install from deb packages
built by our hudson server.

Richard

Hugo Flambó

unread,
Nov 23, 2009, 11:33:18 AM11/23/09
to bigblueb...@googlegroups.com
ok...so that client was the good one...

i've seen something on the list too related to a bug or something
...not sure now

# The start/stop RTP port the application is going to use
# for the media stream.
startRTPPort=3000
stopRTPPort=3029

not sure about the conf i was using...i'll have to check one of this days ...

can this cause be the cause of problems on the app_meetme ??
Nov 11 12:20:07] WARNING[11679]: app_meetme.c:774 build_conf: Unable
>>>>> to open pseudo device


Thanks

BR

Hugo

Richard Alam

unread,
Nov 23, 2009, 11:39:02 AM11/23/09
to bigblueb...@googlegroups.com
It won't cause a problem with app_meetme.

The problem will be that you'll be able to use the voice (headset
icon) initially then after several users you won't be
able to connect.

That's because, the number of users is tied to the RTP port (which
maps to the sip.conf entries). Before, the stopRTP
was 3099 but you only have 3029 users in sip.conf. So when a user
registers 3030 it will fail. Until you reach 3099 which
will then rollback to 3000.

Richard

Walter Tak

unread,
Nov 23, 2009, 11:54:11 AM11/23/09
to bigblueb...@googlegroups.com
I'm not sure if I need dahdi_transcode ; however during my trial to get
things working I stumbled upon that module (probably had errors in my logs
complaining about transcode not being available).

It could be due to the packages I installed (from Debian / apt-get) before I
compiled them from source as mentioned in the guides on
http://code.google.com/p/bigbluebutton/ (install from source Ubuntu or
CentOS).

However I didn't document my actions so I can't recall how it fixed it
exactly. It might also be pretty OS dependent.

Anyway once the dahdi modules got loaded Asterisk worked fine.

I assume you really followed these steps closely ?

**********
# install dahdiwget
http://downloads.asterisk.org/pub/telephony/dahdi-linux-complete/dahdi-linux-complete-2.2.0.2+2.2.0.tar.gztar
xzf dahdi-linux-complete-2.2.0.2+2.2.0.tar.gzcd
dahdi-linux-complete-2.2.0.2+2.2.0# if you get an error about the kernel
sources not being installed# (and you installed kernel-devel like above
says to)# try rebooting your machine - you may have run a yum update that#
updated the kernel version but you are still running on the old#
versionmake allmake installmake configcd ..**********

It's pretty easy to overlook a step. Including the part where you need to
start dahdi manually :)

Anyway the module 'meetme' in Asterisk really depends on Dahdi:

~/bbb/asterisk-1.4.25# grep meetme *

snip---
menuselect-tree:<member name="app_meetme" displayname="MeetMe conference
bridge" remove_on_change="apps/app_meetme.o apps/app_meetme.so">
menuselect-tree: <depend>app_meetme</depend>
menuselect.makedeps:MENUSELECT_DEPENDS_app_meetme=DAHDI
menuselect.makeopts:MENUSELECT_BUILD_DEPS=res_adsi chan_local
res_indications app_meetme res_monitor res_smdi res_features
---

As you can see meetme won't be built if DAHDI isn't there.

Afaik, DAHDI depends on DAHDI_CODEC and DAHDI_CODEC depends on
DAHDI_TRANSCODE if I read the source correctly:

~/bbb/asterisk-1.4.25# grep dahdi *

snip---
menuselect.makedeps:MENUSELECT_DEPENDS_chan_dahdi=RES_SMDI DAHDI TONEZONE
RES_FEATURES PRI
menuselect.makedeps:MENUSELECT_DEPENDS_codec_dahdi=DAHDI_TRANSCODE DAHDI
---

Not sure though. The compilation of Dahdi really needs the kernel-headers or
else it won't compile at etc.

Hope this helps a bit.

W.


----- Original Message -----
From: "Hugo Flamb�" <hfl...@gmail.com>
To: <bigblueb...@googlegroups.com>
Sent: Monday, 23 November 2009 17:00
Subject: Re: [bigbluebutton-dev] got BBB running on Debian / fresh host


Hugo Flambó

unread,
Nov 23, 2009, 12:01:26 PM11/23/09
to bigblueb...@googlegroups.com
yes ...

The difference was that I have used independent modules for dahdi not
the complete package but core + tools...
I'm sure that dahdi was working right...I have compiled with kernel
too...with success...

I'll give it a try again one of this days..and maybe rebuild asterisk
from source again... I guess there's the problem....

Thanks Walter

BR

Hugo

Walter Tak

unread,
Nov 23, 2009, 2:23:27 PM11/23/09
to bigblueb...@googlegroups.com
The packages of dahdi did not work for me. Even if I compiled Asterisk
myself. Both had to be compiled like described in the guide. I'd like to see
packages instead of having to compile things (that's especially "not done"
on production-servers). But then again ; the switch from ztdummy to dahdi
seems to be pretty recent for the Asterisk ppl so BBB and their components
are pretty "under development" and then you can expect changes and unwanted
side-effects like this.

Walter

----- Original Message -----
From: "Hugo Flamb�" <hfl...@gmail.com>
To: <bigblueb...@googlegroups.com>
Sent: Monday, 23 November 2009 18:01
Subject: Re: [bigbluebutton-dev] got BBB running on Debian / fresh host


Hugo Flambó

unread,
Nov 23, 2009, 4:55:20 PM11/23/09
to bigblueb...@googlegroups.com
Yes, I have BBB running on my local laptop.... the intent was to do
the same the same on EC2 but for same reason on VM I got into the
pseudo device problem....That's why I was curious how did you solve
it... I've checked all thing and dig on google to possible causes...

http://www.freepbx.org/support/documentation/faq/common-problems/invalid-conference

Walter Tak

unread,
Nov 23, 2009, 5:46:55 PM11/23/09
to bigblueb...@googlegroups.com
My Red5 console is showing a zillion lines like:

RtpStreamReceiver - run -> encodedBytes = 160, incomingDecodedFrameSize =
160.
RtpStreamReceiver - run -> pkt.length = 172, offset = 12, length = 160.
RtpStreamReceiver - run -> encodedBytes = 160, incomingDecodedFrameSize =
160.
RtpStreamReceiver - run -> pkt.length = 172, offset = 12, length = 160.
RtpStreamReceiver - run -> encodedBytes = 160, incomingDecodedFrameSize =
160.
RtpStreamReceiver - run -> pkt.length = 172, offset = 12, length = 160.
RtpStreamReceiver - run -> encodedBytes = 160, incomingDecodedFrameSize =
160.
RtpStreamReceiver - run -> pkt.length = 172, offset = 12, length = 160.
RtpStreamReceiver - run -> encodedBytes = 160, incomingDecodedFrameSize =
160.
RtpStreamReceiver - run -> pkt.length = 172, offset = 12, length = 160.


How can I turn of these debugging messages to the console ?

Tnx,
Walter

Richard Alam

unread,
Nov 23, 2009, 5:50:50 PM11/23/09
to bigblueb...@googlegroups.com
Hi Walter,

Sorry, I think it can't be turned off from command line or config..you
need to edit the source.

Do you sleep? :)

Richard
> --
>
> You received this message because you are subscribed to the Google Groups "BigBlueButton-dev" group.
> To post to this group, send email to bigblueb...@googlegroups.com.
> To unsubscribe from this group, send email to bigbluebutton-...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/bigbluebutton-dev?hl=.
>
>
>



--

Walter Tak

unread,
Nov 23, 2009, 8:01:16 PM11/23/09
to bigblueb...@googlegroups.com
Hi,

when I click on the "schedule" button on the "homepage" of BBB (see
http://bbb.waltertak.com:81/ for a live demo) I do not end up on that server
but instead nginx
leads me to the internal IP of the server (in this case 192.168.199.2).

(I wrote a full page of details and decided to scrap it to minimize the load
on the Google group servers)

Anyway the firebug output:

GET http://bbb.waltertak.com:81/bigbluebutton 302 moved temporarily
GET http://192.168.199.2/bigbluebutton/ 302 moved temporarily
GET http://192.168.199.2/bigbluebutton/auth/login?targetUri=%2F 200 OK

I can edit the URL and use
http://bbb.waltertak.com:81/bigbluebutton/auth/login?targetUri=%2F but again
it redirects me to URL 192.168.199.2 which works Here, but not There.

Tnx for any advice,
Walter

sawangpong muadphet

unread,
Nov 23, 2009, 10:49:10 PM11/23/09
to bigblueb...@googlegroups.com
Dear Richard
How can I upload PPT format to BBB.
Regards.
Sawangpong

Richard Alam

unread,
Nov 24, 2009, 10:35:14 AM11/24/09
to bigblueb...@googlegroups.com
Check /var/lib/tomcat6/webapps/bigbluebutton/WEB-INF/classes/bigbluebutton.properties

Change bigbluebutton.web.serverURL=http://192.168.0.182 to point to your host.

Or use the command bbb-conf --setip <host>

Richard
> --
>
> You received this message because you are subscribed to the Google Groups "BigBlueButton-dev" group.
> To post to this group, send email to bigblueb...@googlegroups.com.
> To unsubscribe from this group, send email to bigbluebutton-...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/bigbluebutton-dev?hl=en.

Walter Tak

unread,
Nov 23, 2009, 7:12:17 PM11/23/09
to bigblueb...@googlegroups.com
> Hi Walter,
>
> Sorry, I think it can't be turned off from command line or config..you
> need to edit the source.

Ok no problem, where is that nasty critter RtpStreamReceiver located ? Is
that part of Xuggler ?

> Do you sleep? :)

Actually I'm pretty good at that after many years of hard training. However
sleep is overrated obviously when interesting projects are on the horizon ;)

W.

Fred Dixon

unread,
Nov 24, 2009, 11:25:41 AM11/24/09
to BigBlueButton-dev
Hi Walter,

Check these two areas:

http://code.google.com/p/bigbluebutton/wiki/ConfigurationFiles#/var/lib/tomcat/webapps/bigbluebutton/WEB-INF/classes/bigbluebut

check the property bigbluebutton.web.serverURL that it's correct (it
looks good as I just tried clicking Join and it worked).

To edit the session, you are connecting to the Grails application,
which is running on Tomcat, which is listening on port 8080 (by
default). This means, when you connect to http://bbb.waltertak.com:81/bigbluebutton,
nginx redirects this internally to port 8080.

The key configuration file that makes this possible is /etc/nginx/
sites-available/bigbluebutton. Here's the redirect for specification
for URL's that begin with "/bigbluebutton"


location /bigbluebutton {
proxy_pass http://127.0.0.1:8080;
proxy_redirect default;
proxy_set_header X-Forwarded-For
$proxy_add_x_forwarded_for;

client_max_body_size 10m;
client_body_buffer_size 128k;

proxy_connect_timeout 90;
proxy_send_timeout 90;
proxy_read_timeout 90;

proxy_buffer_size 4k;
proxy_buffers 4 32k;
proxy_busy_buffers_size 64k;
proxy_temp_file_write_size 64k;

include fastcgi_params;
}

Note: To enable this configuration file in nginx, you need a symbolic
link from /etc/nginx/sites-enabled/bigbluebutton -> /etc/nginx/sites-
available/bigbluebutton, which you can do with the command

ln -s /etc/nginx/sites-available/bigbluebutton /etc/nginx/sites-
enabled/bigbluebutton

Check that
1. /etc/nginx/sites-available/bigbluebutton has an entry for "/
bigbluebutton"
2. The symbolic link is in place.

If both these look OK, then let us know and we'll dig deeper into what
is causing the redirect.

Regards,... Fred




On Nov 23, 8:01 pm, "Walter Tak" <wal...@waltertak.com> wrote:
> Hi,
>
> when I click on the "schedule" button on the "homepage" of BBB (seehttp://bbb.waltertak.com:81/for a live demo) I do not end up on that server
> but instead nginx
> leads me to the internal IP of the server (in this case 192.168.199.2).
>
> (I wrote a full page of details and decided to scrap it to minimize the load
> on the Google group servers)
>
> Anyway the firebug output:
>
> GEThttp://bbb.waltertak.com:81/bigbluebutton302 moved temporarily
> GEThttp://192.168.199.2/bigbluebutton/302 moved temporarily
> GEThttp://192.168.199.2/bigbluebutton/auth/login?targetUri=%2F200 OK
>
> I can edit the URL and usehttp://bbb.waltertak.com:81/bigbluebutton/auth/login?targetUri=%2Fbut again

Walter Tak

unread,
Nov 24, 2009, 11:52:38 AM11/24/09
to bigblueb...@googlegroups.com
Hi,

That line has been editted already and reads:

[/var/lib/tomcat6/webapps/bigbluebutton/WEB-INF/classes/bigbluebutton.properties]
--
bigbluebutton.web.serverURL=http://bbb.waltertak.com:81
--

I'll try to find the 'redirect' in the Tomcat webapp (if that's the right
place to look) ; I'm currently doing a 3 month BBB101 course in the next 2
days.

Thanks,
Walter

----- Original Message -----

Richard Alam

unread,
Nov 24, 2009, 12:00:48 PM11/24/09
to bigblueb...@googlegroups.com
Try changing /etc/nginx/sites-enabled/bigbluebutton to the following

server {
listen 81;
server_name bbb.waltertak.com;

See if that works.

Richard

Walter

unread,
Nov 24, 2009, 12:07:59 PM11/24/09
to BigBlueButton-dev
The config of nginx (/etc/nginx/sites-available/bigbluebutton) is ok ;
the link on the first page seems not to listen to the config though,
even while the rest of the site does (click the link, everything is
operational, except the admin page).

See http://bbb.waltertak.com:81/

[/etc/nginx/sites-available/bigbluebutton]

...
server {
listen 192.168.199.2:80;
# server_name 192.168.0.136;
# server_name 172.19.3.9;
server_name 192.168.199.2;

....
location /bigbluebutton {
#proxy_pass http://bbb.waltertak.com:8080;
proxy_pass http://172.19.3.9:8080;
proxy_redirect default;
proxy_set_header X-Forwarded-For
$proxy_add_x_forwarded_for;

client_max_body_size 10m;
client_body_buffer_size 128k;

proxy_connect_timeout 90;
proxy_send_timeout 90;
proxy_read_timeout 90;

proxy_buffer_size 4k;
proxy_buffers 4 32k;
proxy_busy_buffers_size 64k;
proxy_temp_file_write_size 64k;

include fastcgi_params;
}

Note the commented lines where I already made corrections ; this
works ; except ... for the redirect to the admin-menu.

There might be an issue with the first [listen] line where I specify
host + port. That's because there are other services running on the
same server on port 80 but which are bound to another IP (yep it's a
pretty complex box ; running mysql 4 and 5 , apache 1 and 2 , bbb ,
nginx , Red5 , tomcat etc :)

However that still doesn't explain why a request for http://server/bigbluebutton
suddenly gets redirected to the original IP of nginx. I'll have a peek
in a source as well once I know where to look.

Thanks,
Walter

Walter

unread,
Nov 24, 2009, 12:18:45 PM11/24/09
to BigBlueButton-dev
Can't do that ; the hostname bbb.waltertak.com resolves to a real-
world IP. The server only has local IP's. Ofcourse that's the problem
but then again ; I rather have all of my servers behind a solid
firewall.

On my (Linux) firewall I use a portforwarding rule (iptables) to
redirect incoming traffic for port 81 to the internal box. I changed
the redirect from port 81 on the firewall to also redirect to port 81
on the BBB-box but that doesn't change the behaviour. (nginx is now
listerning on port 81 as well ofcourse).

It seems the problem is that a request for http://server/bigbluebutton
is not proxy forwarded by nginx to the correct target .

--
location /bigbluebutton {
proxy_pass http://172.19.3.9:8080;
proxy_redirect default;
--

Instead it redirects to http://192.168.199.2:81/bigbluebutton/auth/login?targetUri=%2F

(See http-log in other messages)

This might be a pure nginx problem in combination with my non-so-usual
redirection.

W.
> >> Change bigbluebutton.web.serverURL=http://192.168.0.182to point to your
> >> host.
>
> >> Or use the command bbb-conf --setip <host>
>
> >> Richard
>
> >> On Mon, Nov 23, 2009 at 8:01 PM, Walter Tak <wal...@waltertak.com> wrote:
> >>> Hi,
>
> >>> when I click on the "schedule" button on the "homepage" of BBB (see
> >>>http://bbb.waltertak.com:81/for a live demo) I do not end up on that
> >>> server
> >>> but instead nginx
> >>> leads me to the internal IP of the server (in this case 192.168.199.2).
>
> >>> (I wrote a full page of details and decided to scrap it to minimize the
> >>> load
> >>> on the Google group servers)
>
> >>> Anyway the firebug output:
>
> >>> GEThttp://bbb.waltertak.com:81/bigbluebutton302 moved temporarily
> >>> GEThttp://192.168.199.2/bigbluebutton/302 moved temporarily
> >>> GEThttp://192.168.199.2/bigbluebutton/auth/login?targetUri=%2F200 OK

Walter

unread,
Nov 24, 2009, 12:58:31 PM11/24/09
to BigBlueButton-dev
What script/source is actually handling requests for URL /
bigbluebutton/ ?

I understand that nginx redirects http-requests for /bigbluebutton to
tomcat on port 8080 and following that path I think requests are being
handled by

[ /bigbluebutton-read-only/bigbluebutton-web/grails-app/controllers/
org/bigbluebutton/web/controllers/PortalController.groovy ] and
[ /bigbluebutton-read-only/bigbluebutton-web/grails-app/controllers/
org/bigbluebutton/web/controllers/AuthController.groovy ] right ?

I * think* some redirecting goes wrong in AuthController.groovy at
lines:

---
def index = { redirect(action: 'login', params: params) }

def login = {
return [ username: params.username, rememberMe:
(params.rememberMe != null), targetUri: params.targetUri ]
}
---

Need to find out how to recompile the tomcat-app. It's too much
guessing right now.

W.

Richard Alam

unread,
Nov 24, 2009, 1:09:56 PM11/24/09
to bigblueb...@googlegroups.com
To compile, type "grails prod war" in your bbb-web source.

Checkout the "redirect" api from http://grails.org/doc/latest/ under
controllers. Hopefully, you can
hardcode the URL.

Haven't had the same setup that you have (having BBB behind the
firewall), so I guess that's why we haven't ran into this.

Richard

Walter

unread,
Nov 24, 2009, 1:22:01 PM11/24/09
to BigBlueButton-dev
Tnx,

Was already in the process of downloading Grails.

I'm fully aware that exotic network setups can never be fully tested
(nor should you want to do that).

I could try to install nginx on my firewall. In the end I do like to
create dedicated VMs for all components anyway (tomcat, red5 and
asterisk) for easier maintenance and easier loadbalancing.

Walter

unread,
Nov 24, 2009, 1:41:36 PM11/24/09
to BigBlueButton-dev
Ok that fixed it.

Everything now works like intended.

I added nginx to my firewall, added proxy_passes also for the / (root)
and /client content and let nginx listen on port 81 on the firewall,
instead of redirecting port 81 to the bbb-box behind it.

See for a quick test http://bbb.waltertak.com:81/

I changed the admin login but added a new user "b...@test.com" pass
"bobo" ; should work.

Glad that I don't need to recompile things (yet).

Next step ; integration of our frontend (and users-database) to remove
the current 'homepage'. I already added a few conferences and sessions
by directly editting records in MySQL but I read there is an API
lurking around the corner ?

I'd rather not do direct queries on the same database BBB uses, that
would be kinda dirty and very version dedepent.

Tips ?
W.

Richard Alam

unread,
Nov 24, 2009, 1:49:10 PM11/24/09
to bigblueb...@googlegroups.com

Walter Tak

unread,
Nov 24, 2009, 3:16:55 PM11/24/09
to bigblueb...@googlegroups.com
Ok great. API works fine for creating and automated joining of a conference.
Good.

There is supposed to be support for a checksum based on a secret that's only
to be shared between BBB and the calling server (never going over the wire)
however I'm able to call the API without any form of authentication or
autorization is that correct ?

e.g. request [
/bigbluebutton/api/create?name=WalterAPI&meetingID=abcabc&attendeePW=ap&moderatorPW=mp
] works as intended and returns a valid xml response.

Is security in the API a future feature or should it be enabled in a
config-file ?

Tnx,
Walter

----- Original Message -----
From: "Richard Alam" <ritz...@gmail.com>
To: <bigblueb...@googlegroups.com>
Sent: Tuesday, 24 November 2009 19:49
Subject: Re: [bigbluebutton-dev] Re: URL to admin-page does not follow
host-settings ?


DJP

unread,
Nov 24, 2009, 3:39:23 PM11/24/09
to BigBlueButton-dev
Hi Walter,

The API is still under development by Jeremy so that's why the
SECURITY_SALT isn't in use for now and does'nt work... ( Cf code
http://code.google.com/p/bigbluebutton/source/browse/trunk/bigbluebutton-web/grails-app/controllers/org/bigbluebutton/web/controllers/ApiController.groovy
).

Since it's planned in the current iteration (BBB 0.63), itshould be
available on mid-december...

Cya,
On 24 nov, 21:16, "Walter Tak" <wal...@waltertak.com> wrote:
> Ok great. API works fine for creating and automated joining of a conference.
> Good.
>
> There is supposed to be support for a checksum based on a secret that's only
> to be shared between BBB and the calling server (never going over the wire)
> however I'm able to call the API without any form of authentication or
> autorization is that correct ?
>
> e.g. request [
> /bigbluebutton/api/create?name=WalterAPI&meetingID=abcabc&attendeePW=ap&mod eratorPW=mp
>  ] works as intended and returns a valid xml response.
>
> Is security in the API a future feature or should it be enabled in a
> config-file ?
>
> Tnx,
> Walter
>
>
>
> ----- Original Message -----
> From: "Richard Alam" <ritza...@gmail.com>
> To: <bigblueb...@googlegroups.com>
> Sent: Tuesday, 24 November 2009 19:49
> Subject: Re: [bigbluebutton-dev] Re: URL to admin-page does not follow
>
> host-settings ?
>
> Jeremy created an API documented herehttp://bigbluebutton.googlecode.com/files/bbb-xml-api.pdf
>
> You can take a look at what's implemented herehttp://code.google.com/p/bigbluebutton/source/browse/trunk/bigbluebut...
>
> Richard
>
> On Tue, Nov 24, 2009 at 1:41 PM, Walter <wal...@waltertak.com> wrote:
> > Ok that fixed it.
>
> > Everything now works like intended.
>
> > I added nginx to my firewall, added proxy_passes also for the / (root)
> > and /client content and let nginx listen on port 81 on the firewall,
> > instead of redirecting port 81 to the bbb-box behind it.
>
> > See for a quick testhttp://bbb.waltertak.com:81/
> BigBlueButtonhttp://www.bigbluebutton.orghttp://code.google.com/p/bigbluebutton

Richard Alam

unread,
Nov 24, 2009, 3:51:50 PM11/24/09
to bigblueb...@googlegroups.com
The securitySalt is null which makes it disabled by default.

String securitySalt = null;

Try setting it in bigbluebutton.properties

beans.apiController.securitySalt = 639259d4-9dd8-4b25-bf01-95f9567eaf4b

Then restart tomcat.

I haven't tried this yet but I think that's how it needs to be set.

Richard

Jeremy Thomerson

unread,
Nov 24, 2009, 4:31:36 PM11/24/09
to bigblueb...@googlegroups.com
Yes - if you set it like this, security is enabled.  Let me know if you have any issues.  I'm out of the country on vacation but will have email access for about another 24 hours.  Then you're on your own for a couple weeks - I'll be where I'm really unavailable.  :)
Reply all
Reply to author
Forward
0 new messages