| Summary: | ASTERISK-05015: Crashes with nocdr after resetcdr in dial plan | ||
| Reporter: | xrobau (xrobau) | Labels: | |
| Date Opened: | 2005-09-07 07:53:49 | Date Closed: | 2005-10-28 13:15:11 |
| Priority: | Critical | Regression? | No |
| Status: | Closed/Complete | Components: | Core/General |
| Versions: | Frequency of Occurrence | ||
| Related Issues: | |||
| Environment: | Attachments: | ( 0) cdrcrash.txt | |
| Description: | See xrobaus note further down. | ||
| Comments: | By: Olle Johansson (oej) 2005-09-07 07:54:44 Hey, I could make a copy! By: Tilghman Lesher (tilghman) 2005-09-08 16:39:29 Could we replicate the relevant backtrace over to this bug, too? By: xrobau (xrobau) 2005-09-09 04:37:08 Sorry for the delay in attaching this file. It's _very_ easily duplicateable. Establish a call (from phone1) to phone2. Put it on hold. Establish another call (to phone3). Transfer the second call (phone3) to the first call (phone2). Hang up either of the phones, and, in your extension plan, have this: [macro-hangupcall] exten => s,1,ResetCDR(w) exten => s,2,NoCDR() exten => s,3,Wait(5) exten => s,4,Hangup (That's an AMP-ism, for those that don't recongnise it). The asterisk box will crash. By: Mark Spencer (markster) 2005-09-12 00:41:14 I can't seem to duplicate this, I'll need to ssh into your machine and watch it happen from there. Find me on IRC (kram, irc.freenode.net, #asterisk) By: Mark Spencer (markster) 2005-10-12 02:13:20 Tried again at astricon to duplicate it and were unable to. We either need ssh access in the next day or two or we're going to have to close this one out as suspended. By: xrobau (xrobau) 2005-10-12 02:18:35 This was repeatable with some sip code that oej provided me with - I was under the impression that he was going to submit it 'real soon now' when this bug was created, but it hasn't happened yet. There definately _is_ a problem with CDR's, A@H/AMP uses ResetCDR/NoCDR on macro-hangup, which is called on every h event. At the moment, it's not reliable, and it will crash after n calls, but n ranges from 5 to >2000. The crash occurs in different places, and it's much easier to trigger when doing transfers on a Snom phone. Olle, any comments? By: twisted (twisted) 2005-10-17 23:13:02 so this bug is not producable with current CVS HEAD? It sounds like you are running modified code. If the code you are running differs from CVS HEAD, since it's not reproducable, we'll have to close this out. By: Olle Johansson (oej) 2005-10-18 12:18:30 Using standard SIP channel or my transfer patch did not make a difference, did it Xrobau? By: Mark Spencer (markster) 2005-10-28 13:15:02 Already fixed in CVS head. | ||