headers
Anton | 15 Aug 14:30

ZAP channel hang

Just tired from one longstanding bug: accidently (once in a month or so) some ZAP channel got locked (there
is power on the ZAP line but no tone)
you may find below 2 of the channels having over 4 hours of the uptime, corresponding ZAP extension on
channel bank - 
is just silent - just until asterisk (not zaptel) restart. This bug remains since asterisk v 1.2. I've been
filling a 
bug report a long ago. If someone of the developers willing to address this issue, 
I just need to know what info should I gather from my working system on the next reoccurence of this case, 
and will fill a bugreport with all of the needed details.

Asterisk 1.4.21.1 
Zaptel 1.4.11

CLI> show channels verbose
Channel              Context              Extension        Prio State   Application  Data                      CallerID        Duration Accountcode BridgedTo
Zap/21-1             internal                                 1 Up      Bridged Call Zap/27-1                  127             04:49:01             Zap/27-1
Zap/27-1             internal             116                 3 Up      Dial         Zap/21r2|25|thWwoj        127             04:49:08             Zap/21-1
2 active channels
1 active call

Anton.

_______________________________________________
--Bandwidth and Colocation Provided by http://www.api-digital.com--

AstriCon 2008 - September 22 - 25 Phoenix, Arizona
Register Now: http://www.astricon.net

asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
(Continue reading)

Permalink | Reply |
headers
Tilghman Lesher | 15 Aug 16:01

Re: ZAP channel hang

On Friday 15 August 2008 07:34:09 Anton wrote:
> Just tired from one longstanding bug: accidently (once in a month or so)
> some ZAP channel got locked (there is power on the ZAP line but no tone)
> you may find below 2 of the channels having over 4 hours of the uptime,
> corresponding ZAP extension on channel bank - is just silent - just until
> asterisk (not zaptel) restart. This bug remains since asterisk v 1.2. I've
> been filling a bug report a long ago. If someone of the developers willing
> to address this issue, I just need to know what info should I gather from
> my working system on the next reoccurence of this case, and will fill a
> bugreport with all of the needed details.
>
> Asterisk 1.4.21.1
> Zaptel 1.4.11
>
> CLI> show channels verbose
> Channel              Context              Extension        Prio State  
> Application  Data                      CallerID        Duration Accountcode
> BridgedTo Zap/21-1             internal                                 1
> Up      Bridged Call Zap/27-1                  127             04:49:01    
>         Zap/27-1 Zap/27-1             internal             116             
>    3 Up      Dial         Zap/21r2|25|thWwoj        127            
> 04:49:08             Zap/21-1 2 active channels
> 1 active call

First, it would probably be good to ensure this is not a deadlock in Asterisk
of some sort.  Recompile with DEBUG_THREADS, and then, when this problem
occurs, run 'core show locks' and show the result.

--

-- 
Tilghman
(Continue reading)

Permalink | Reply |
headers
Anton | 15 Aug 17:35

Re: ZAP channel hang

Done, waiting for reoccurence...

On Friday 15 August 2008 19:01, Tilghman Lesher wrote:
> On Friday 15 August 2008 07:34:09 Anton wrote:
> > Just tired from one longstanding bug: accidently (once
> > in a month or so) some ZAP channel got locked (there is
> > power on the ZAP line but no tone) you may find below 2
> > of the channels having over 4 hours of the uptime,
> > corresponding ZAP extension on channel bank - is just
> > silent - just until asterisk (not zaptel) restart. This
> > bug remains since asterisk v 1.2. I've been filling a
> > bug report a long ago. If someone of the developers
> > willing to address this issue, I just need to know what
> > info should I gather from my working system on the next
> > reoccurence of this case, and will fill a bugreport
> > with all of the needed details.
> >
> > Asterisk 1.4.21.1
> > Zaptel 1.4.11
> >
> > CLI> show channels verbose
> > Channel              Context              Extension    
> >    Prio State Application  Data                     
> > CallerID        Duration Accountcode BridgedTo Zap/21-1
> >             internal                                 1
> > Up      Bridged Call Zap/27-1                  127     
> >        04:49:01 Zap/27-1 Zap/27-1             internal 
> >            116 3 Up      Dial        
> > Zap/21r2|25|thWwoj        127 04:49:08            
> > Zap/21-1 2 active channels 1 active call
(Continue reading)

Permalink | Reply |
headers
Paradise Dove | 15 Aug 22:21

Re: ZAP channel hang

Hi,
I have the same problem and i'm sure that it's not a deadlock and a
simple "soft hangup" on one of the hung channels will kick them off!
i can reproduce  it easily,
one way is to call your box from outside on fxo channels and make your
scripts to run DISA application and give you an extension tone with
outside dialing permission, now just dial a number using another fxo
channel. with this scenario in more than 80% of times after ending the
conversation conversation, both bridged fxo channels will remain
connected for ever!!
the other way is to call a number via fxo channels and then hold the
call and call another number via fxo channels and then bridge both
calls together (transfer). also in this scenario the bridged fxo
channels more than 60% of times will remain hung.
the same problem exists if you bridge (transfer) two incoming calls
from fxo channels.
i have no idea how to fix this problem!? :(

i hope this help,
thanks
paradise

On Fri, Aug 15, 2008 at 7:05 PM, Anton <anton.vazir <at> gmail.com> wrote:
> Done, waiting for reoccurence...
>
> On Friday 15 August 2008 19:01, Tilghman Lesher wrote:
>> On Friday 15 August 2008 07:34:09 Anton wrote:
>> > Just tired from one longstanding bug: accidently (once
>> > in a month or so) some ZAP channel got locked (there is
>> > power on the ZAP line but no tone) you may find below 2
(Continue reading)

Permalink | Reply |
headers
Tilghman Lesher | 15 Aug 23:33

Re: ZAP channel hang

On Friday 15 August 2008 15:21:04 Paradise Dove wrote:
> I have the same problem and i'm sure that it's not a deadlock and a
> simple "soft hangup" on one of the hung channels will kick them off!
> i can reproduce  it easily,
> one way is to call your box from outside on fxo channels and make your
> scripts to run DISA application and give you an extension tone with
> outside dialing permission, now just dial a number using another fxo
> channel. with this scenario in more than 80% of times after ending the
> conversation conversation, both bridged fxo channels will remain
> connected for ever!!
> the other way is to call a number via fxo channels and then hold the
> call and call another number via fxo channels and then bridge both
> calls together (transfer). also in this scenario the bridged fxo
> channels more than 60% of times will remain hung.
> the same problem exists if you bridge (transfer) two incoming calls
> from fxo channels.

If that is the case, then I believe there is no good way to detect this
problem.  The issue is that the signalling for Loopstart FXO channels
is deficient.  This is normally not a problem, because FXO channels
are rarely bridged with other FXO channels.  Instead, they are usually
bridged with either FXS channels (where the hangup signal is easily
detectable) or SIP or other digital channels, where the hangup signal
is digital in nature and is thus not subject to this problem.

You could certainly try callprogress detection, which attempts to do
a better job of detecting hangups, busy signals, and other exceptional
conditions; however, callprogress is not perfect and may lead to early
termination of calls in certain situations.  The only real solution is to
upgrade to a digitally signalled line (BRI, PRI, or SIP trunking).
(Continue reading)

Permalink | Reply |
headers
Anton | 16 Aug 10:15

Re: ZAP channel hang

My PBX work just as bridging FXS to FXS or SIP to FXS or IAX 
to FXS (depends from origin of the call) - T1 card 
connected to the channel bank and second CB is XORCOMM USB 
CB. extensions are configured as FXOKS and connected to the 
regular prones for the regular office work (mostly call 
transfer, call pickup, no conferences and no DISA, no 
meetme) - everything is very basic. Looks CB can reliably 
detect phone hangup, but is it's the case than asterisk 
cannot handle it sometimes? Anyway next reoccurence will 
show if there is any possible lock or no. BTW - the same 
situation persist among any zaptel device - XORCOMM USB 
ChannelBank behave exactly the same as the pair of 
T1->Rhino T1-ChannelBank, so it's equipment independant, 
just it must me a zaptel device. 

On Saturday 16 August 2008 01:21, Paradise Dove wrote:
> Hi,
> I have the same problem and i'm sure that it's not a
> deadlock and a simple "soft hangup" on one of the hung
> channels will kick them off! i can reproduce  it easily,
> one way is to call your box from outside on fxo channels
> and make your scripts to run DISA application and give
> you an extension tone with outside dialing permission,
> now just dial a number using another fxo channel. with
> this scenario in more than 80% of times after ending the
> conversation conversation, both bridged fxo channels will
> remain connected for ever!!
> the other way is to call a number via fxo channels and
> then hold the call and call another number via fxo
> channels and then bridge both calls together (transfer).
(Continue reading)

Permalink | Reply |
headers
Anton | 25 Sep 09:24

Re: ZAP channel hang

Hi Dove,

Did you found any workarround?

On Saturday 16 August 2008 01:21, Paradise Dove wrote:
> Hi,
> I have the same problem and i'm sure that it's not a
> deadlock and a simple "soft hangup" on one of the hung
> channels will kick them off! i can reproduce  it easily,
> one way is to call your box from outside on fxo channels
> and make your scripts to run DISA application and give
> you an extension tone with outside dialing permission,
> now just dial a number using another fxo channel. with
> this scenario in more than 80% of times after ending the
> conversation conversation, both bridged fxo channels will
> remain connected for ever!!
> the other way is to call a number via fxo channels and
> then hold the call and call another number via fxo
> channels and then bridge both calls together (transfer).
> also in this scenario the bridged fxo channels more than
> 60% of times will remain hung. the same problem exists if
> you bridge (transfer) two incoming calls from fxo
> channels.
> i have no idea how to fix this problem!? :(
>
> i hope this help,
> thanks
> paradise
>
> On Fri, Aug 15, 2008 at 7:05 PM, Anton
(Continue reading)

Permalink | Reply |
headers
Anton | 22 Aug 12:52

Re: ZAP channel hang

Reoccured. No any locks. Just 2 silent channels again....

On Friday 15 August 2008 19:01, Tilghman Lesher wrote:
> On Friday 15 August 2008 07:34:09 Anton wrote:
> > Just tired from one longstanding bug: accidently (once
> > in a month or so) some ZAP channel got locked (there is
> > power on the ZAP line but no tone) you may find below 2
> > of the channels having over 4 hours of the uptime,
> > corresponding ZAP extension on channel bank - is just
> > silent - just until asterisk (not zaptel) restart. This
> > bug remains since asterisk v 1.2. I've been filling a
> > bug report a long ago. If someone of the developers
> > willing to address this issue, I just need to know what
> > info should I gather from my working system on the next
> > reoccurence of this case, and will fill a bugreport
> > with all of the needed details.
> >
> > Asterisk 1.4.21.1
> > Zaptel 1.4.11
> >
> > CLI> show channels verbose
> > Channel              Context              Extension    
> >    Prio State Application  Data                     
> > CallerID        Duration Accountcode BridgedTo Zap/21-1
> >             internal                                 1
> > Up      Bridged Call Zap/27-1                  127     
> >        04:49:01 Zap/27-1 Zap/27-1             internal 
> >            116 3 Up      Dial        
> > Zap/21r2|25|thWwoj        127 04:49:08            
> > Zap/21-1 2 active channels 1 active call
(Continue reading)

Permalink | Reply |
headers
Tzafrir Cohen | 25 Sep 10:07
Favicon
Gravatar

Re: ZAP channel hang

On Fri, Aug 15, 2008 at 09:01:12AM -0500, Tilghman Lesher wrote:

> First, it would probably be good to ensure this is not a deadlock in Asterisk
> of some sort.  Recompile with DEBUG_THREADS, and then, when this problem
> occurs, run 'core show locks' and show the result.

As this thread has risen from the dead, I'll add a comment:

I recently needed to debug some deadlocks in chan_zap. I followed this
advice. Some threads were deadlocked. So I ran 'core show locks'. And
that command happily hung :-(

The thing that did help me to debug things was DETECT_DEADLOCKS . I
assume it has a noticable run-time overhead, and therefore may not be
always a good option in production. But it did help me there.

--

-- 
               Tzafrir Cohen
icq#16849755              jabber:tzafrir.cohen <at> xorcom.com
+972-50-7952406           mailto:tzafrir.cohen <at> xorcom.com
http://www.xorcom.com  iax:guest <at> local.xorcom.com/tzafrir

_______________________________________________
--Bandwidth and Colocation Provided by http://www.api-digital.com--

AstriCon 2008 - September 22 - 25 Phoenix, Arizona
Register Now: http://www.astricon.net

asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
(Continue reading)

Permalink | Reply |

Gmane