[asterisk-bugs] [JIRA] (ASTERISK-24667) IAX hits maxcallnumber limit of 2048 after many calls and don't recover after hangup
Y Ateya (JIRA)
noreply at issues.asterisk.org
Tue Jan 6 17:07:34 CST 2015
[ https://issues.asterisk.org/jira/browse/ASTERISK-24667?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Y Ateya updated ASTERISK-24667:
-------------------------------
Description:
NOTE: This might be clone of ASTERISK-24478 .
I saw the same problem as ASTERISK-24478 and I was able to reproduce it. My setup is as follows:
SIP1 <> IAX Trunk <> SIP2 <> GSM gateway
Hint: Calls always flow from SIP1 (always the caller) to SIP2 (always the callee).
The problem occurs when there is a failure in GSM gateway; which leads to rejecting many of SIP calls (rejected from GSM Gateway). Here is what should happen for increment, schedule decrement and actual decrement of the call number counter:
call comes
increment counter
call rejected
schedule counter decrement after 60 seconds
and this pattern will be repeated until GSM return to normal operation. So after 60 seconds of GSM-back-to-life; all calls will continue as usual. But this DON'T always happen.
What happens occasionally is:
increment, schedule decrement, after 60 seconds decrement event occurs.
increment, schedule decrement, after 60 seconds decrement event occurs.
.
.
increment, schedule decrement, after 60 seconds decrement event occurs.
=======And after a while (~15 minutes in my case), decrement event stops!
increment, schedule decrement
increment, schedule decrement
increment, schedule decrement
Decrement don't happen even after I stopped SIP1 calls.
And the result is that IAX peers have no active calls and max call number is 2048. Raising max call number doesn't solve the problem; Neither decreasing MIN_REUSE_TIME.
The only solution then is to restart asterisk; reload didn't affect call number count.
The attached log is from Jan 1 18:12:13 to Jan 1 18:32:46. You will that the latest decrement "ip callno count decremented" was at Jan 1 18:23:32. For TEN minutes, scheduled decrements didn't happen.
Attachments:
maxcallnumber_count.log.bz2: Extracted counter-related messages from debug log.
maxcallnumber_full.log.bz2: The full log
was:
[Edit by Rusty - Even if we can't track down the issue here, we should improve iax.conf to mention the 2048 default limit and describe the max values for maxcallnumbers and maxcallnumbers_nonvalidated]
Hi,
With ARI, I create channels to make outbounds call via a IAX Trunk. After some time, i have on the asterisk cli :
[Nov 1 16:14:46] ERROR[5236][C-00005077]: chan_iax2.c:2588 peercnt_add: maxcallnumber limit of 2048 for 10.100.5.1 has been reached!
[Nov 1 16:14:46] WARNING[5236][C-00005077]: chan_iax2.c:12400 iax2_request: Unable to create call
[Nov 1 16:14:46] WARNING[5236][C-00005077]: app_dial.c:2421 dial_exec_full: Unable to create channel of type 'IAX2' (cause 34 - Circuit/channel congestion)
There is no maxcallnumber, on each side of the iax Trunk.
core show channels:
Channel Location State Application(Data)
IAX2/autodial_b-2918 (None) Ringing AppDial((Outgoing Line))
Local/0964158794 at REA (None) Ring Dial(IAX2/autodial_b/096415879
1 active channel
1 active call
20672 calls processed
The active call does not exist in reality
module show like iax;
Module Description Use Count Status
chan_iax2.so Inter Asterisk eXchange (Ver 2) 1 Running
1 modules loaded
the asterisk cli stop working and i have to lauch it again.
Environment:
Running Asterisk 13.1.0 compiled from sources and using PJPROJECT.
Ubuntu 12.04 x86_64
was:Ubuntu 12.04 X86_64
Frequency of Occurrence: Occasional (was: Constant)
> IAX hits maxcallnumber limit of 2048 after many calls and don't recover after hangup
> ------------------------------------------------------------------------------------
>
> Key: ASTERISK-24667
> URL: https://issues.asterisk.org/jira/browse/ASTERISK-24667
> Project: Asterisk
> Issue Type: Bug
> Security Level: None
> Components: Channels/chan_iax2
> Affects Versions: 13.1.0
> Environment: Running Asterisk 13.1.0 compiled from sources and using PJPROJECT.
> Ubuntu 12.04 x86_64
> Reporter: Y Ateya
> Severity: Critical
> Attachments: maxcallnumber_count.log.bz2, maxcallnumber_full.log.bz2
>
>
> NOTE: This might be clone of ASTERISK-24478 .
> I saw the same problem as ASTERISK-24478 and I was able to reproduce it. My setup is as follows:
> SIP1 <> IAX Trunk <> SIP2 <> GSM gateway
> Hint: Calls always flow from SIP1 (always the caller) to SIP2 (always the callee).
> The problem occurs when there is a failure in GSM gateway; which leads to rejecting many of SIP calls (rejected from GSM Gateway). Here is what should happen for increment, schedule decrement and actual decrement of the call number counter:
> call comes
> increment counter
> call rejected
> schedule counter decrement after 60 seconds
> and this pattern will be repeated until GSM return to normal operation. So after 60 seconds of GSM-back-to-life; all calls will continue as usual. But this DON'T always happen.
> What happens occasionally is:
> increment, schedule decrement, after 60 seconds decrement event occurs.
> increment, schedule decrement, after 60 seconds decrement event occurs.
> .
> .
> increment, schedule decrement, after 60 seconds decrement event occurs.
> =======And after a while (~15 minutes in my case), decrement event stops!
> increment, schedule decrement
> increment, schedule decrement
> increment, schedule decrement
> Decrement don't happen even after I stopped SIP1 calls.
> And the result is that IAX peers have no active calls and max call number is 2048. Raising max call number doesn't solve the problem; Neither decreasing MIN_REUSE_TIME.
> The only solution then is to restart asterisk; reload didn't affect call number count.
> The attached log is from Jan 1 18:12:13 to Jan 1 18:32:46. You will that the latest decrement "ip callno count decremented" was at Jan 1 18:23:32. For TEN minutes, scheduled decrements didn't happen.
> Attachments:
> maxcallnumber_count.log.bz2: Extracted counter-related messages from debug log.
> maxcallnumber_full.log.bz2: The full log
--
This message was sent by Atlassian JIRA
(v6.2#6252)
More information about the asterisk-bugs
mailing list