OpenVZ Forum: Support » Asterisk problem reported.

Home » General » Support » Asterisk problem reported. (Exsiting Asterisk VE affected by creation of new Asterisk VE)

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Asterisk problem reported. [message #38366]

Fri, 11 December 2009 22:53

unxs
Messages: 21
Registered: September 2009
Location: Oregon, USA

Junior Member

Reported by third party.

Asterisk containers use a shared HN timing device:

From typical Asterisk /etc/vz/conf file
...
DEVNODES="dahdi/pseudo:rw dahdi/ctl:rw dahdi/channel:rw dahdi/timer:rw "
...

Third party reports that running VE PBX drops calls when another Asterisk container is created.

uname -a output of hardware node:

Linux fs104.NNN.com 2.6.18-128.2.1.el5.028stab064.4 #1 SMP Mon Jul 27 12:45:01 MSD 2009 x86_64 x86_64 x86_64 GNU/Linux

Anybody ever heard of this problem?

Report message to a moderator

Re: Asterisk problem reported. [message #38373 is a reply to message #38366]

Sat, 12 December 2009 00:49

mustardman
Messages: 91
Registered: October 2009

Member

Are they running 32bit or 64bit VPS?

Dahdi/Zaptel will not work between 32 and 64bit VPS/Node or visa versa.

I know that does not directly answer your question but it may be related.

[Updated on: Sat, 12 December 2009 00:50]

Report message to a moderator

Re: Asterisk problem reported. [message #38374 is a reply to message #38373]

Sat, 12 December 2009 02:20

unxs
Messages: 21
Registered: September 2009
Location: Oregon, USA

Junior Member

Thanks for your input.

The DAHDI 32/64 problem is well known and the HNs we are talking about have had dozens of production PBX VEs running for quite some time now. They even live migrate between same datacenter HNs during ongoing calls that are not dropped -only a very short audio loss. We are using our alpha unxsVZ container cluster manager.

Of course the meetme Asterisk component (or something called like that) for conference calling may have something to do with this. I no little about the Asterisk system itself. But since the new container creation supposedly crashed a conference call.

May be some fluke SIP trunk/CLEC provider snafu.

May not even be OpenVZ related or dummy dahdi device related at all.

I will try to reproduce the problem in a controlled setting.

Thanks again!

Report message to a moderator

Re: Asterisk problem reported. [message #38375 is a reply to message #38374]

Sat, 12 December 2009 08:16

mustardman
Messages: 91
Registered: October 2009

Member

I can see it crashing currently running conference calls. You should only create new VE/dahdi associations when there are no conference calls going on. Late in the evening for example.

Report message to a moderator

Re: Asterisk problem reported. [message #38377 is a reply to message #38375]

Sat, 12 December 2009 12:32

unxs
Messages: 21
Registered: September 2009
Location: Oregon, USA

Junior Member

Interesting.

If you have any links or other info about the technical aspects of why this happens that would be great. I would like to know more about the technical details.

Since my problem involves creating 100's of Asterisk containers per day (on many clustered HNs.) I need to solve this problem.

The solution as I see it now, is to monitor all Asterisk containers via the autonomic VZ VE management system we use and defer creation (keep in job queue) until any conference call is over (scope is per single HN.) Another option is for the system to autonomously deploy the new Asterisk VE on another cluster member HN that has no ongoing conference calls. In order for this method to work there has to be a way to detect ongoing conference calls (or even better, about to commence conference calls.)

I'm sure there is some device information in the Asterisk VE itself or other Asterisk log file, monitoring system, etc. that can be used to detect ongoing conference calls. Will have to talk to an Asterisk engineer about this.

Thanks for your reply!

Report message to a moderator

Re: Asterisk problem reported. [message #38378 is a reply to message #38377]

Sat, 12 December 2009 20:31

mustardman
Messages: 91
Registered: October 2009

Member

If your doing that many I would assume you have some business clout. I would suggest getting in touch with Asterisk directly or if you have some skilled coders around, dig into the source code.

While your at it, how about doing me a favor and getting the 64/32 problem fixed. Cool

Report message to a moderator

Re: Asterisk problem reported. [message #38380 is a reply to message #38378]

Sat, 12 December 2009 23:18

unxs
Messages: 21
Registered: September 2009
Location: Oregon, USA

Junior Member

lol, yes you are absolutely right. Razz

I will ask these people to get Digium in on this ASAP.

Will ask about the 32/64 issue also. But of course by going 64/64 everything is cool and stable (knock on wood.)

Thx for your input!

Report message to a moderator

Previous Topic:	help ( VM creation has failed)
Next Topic:	About the network interface for OpenVZ

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Thu Jul 18 19:27:51 GMT 2024

Total time taken to generate the page: 0.02546 seconds