Summary: | ASTERISK-14408: crash / call forwarding possibly due to loop detection - 1.4.26rc4 | ||
Reporter: | Alan Graham (zerohalo) | Labels: | |
Date Opened: | 2009-07-01 13:41:30 | Date Closed: | 2010-07-26 21:44:22 |
Priority: | Critical | Regression? | No |
Status: | Closed/Complete | Components: | Core/Channels |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ( 0) bt_scrubbed.txt ( 1) bt.txt ( 2) cli.txt | |
Description: | I'm not sure how to interpret this one. It appears that a customer PBX was forwarded to itself from an outside number, causing a crash. CLI right up to the crash: [Jul 1 12:33:19] VERBOSE[16909] logger.c: -- Executing [s@CONTEXT_REMOVED:22] Dial("SIP/gateway1-b66b22e0", "SIP/5555555555|70|r") in new stack [Jul 1 12:33:19] VERBOSE[10250] logger.c: -- Got SIP response 482 "Loop Detected" back from XX.XXX.XXX.XX [Jul 1 12:33:19] VERBOSE[16910] logger.c: -- Now forwarding SIP/5555555555-b637f1a8 to 'Local/3333333333@inbound' (thanks to SIP/server1-08f125c0) [Jul 1 12:33:19] VERBOSE[16909] logger.c: -- Called 5555555555 [Jul 1 12:33:19] VERBOSE[10250] logger.c: -- Got SIP response 482 "Loop Detected" back from XX.XXX.XXX.XX [Jul 1 12:33:19] WARNING[16910] channel.c: PBX may not have been terminated properly on 'SIP/server1-08f125c0' [Jul 1 12:33:19] VERBOSE[16912] logger.c: -- Executing AppendCDRUserField("Local/3333333333@inbound-2a8e,2", "DST=3333333333|") [Jul 1 12:33:19] WARNING[16911] channel.c: Unable to write to alert pipe on (qlen = 0): Bad file descriptor! [Jul 1 12:33:19] WARNING[16911] channel.c: Unable to write to alert pipe on (qlen = 1): Bad file descriptor! [Jul 1 12:33:19] VERBOSE[16911] logger.c: -- Executing AppendCDRUserField("", "D=3333333333|") [Jul 1 12:33:19] WARNING[10548] channel.c: read() failed: Bad file descriptor [Jul 1 12:33:19] WARNING[10548] channel.c: read() failed: Is a directory [Jul 1 12:33:19] WARNING[10548] channel.c: read() failed: Bad file descriptor backtrace attached | ||
Comments: | By: Alan Graham (zerohalo) 2009-08-11 10:28:11 Second crash, exact same BT and CLI output - is there anything more needed to get this looked at? By: Alan Graham (zerohalo) 2009-08-11 16:14:28 Persists in 1.4.26 By: Matthew Nicholson (mnicholson) 2009-09-14 08:57:15 Can you post more detailed instructions for reproducing this bug? I have tried simply having an asterisk system call it self, which did not crash the system. Also please upload a backtrace generated with 'bt full' if possible. By: Matthew Nicholson (mnicholson) 2009-09-24 10:45:54 I am closing this due to lack of feedback. If you can provide the information I have requested, reopen the bug. By: Alan Graham (zerohalo) 2009-10-08 14:18:22 It appears that this happens randomly when a Local/ channel calls an extension which calls another Local/ channel which calls a third Local/ channel on the same box (and within the same context). I cannot reliably reproduce this. 'bt full' forthcoming. By: Alan Graham (zerohalo) 2009-10-08 14:24:45 bt full, thread apply all bt attached. By: Alan Graham (zerohalo) 2009-10-08 14:45:06 attached is what I got from the messages log from a second crash that happened within ten minutes of the last one. The files referenced do exist - This machine is running 1.4.26.2 w/ filestream_frisolate-1.4.diff2.txt from ASTERISK-1543609. By: Alan Graham (zerohalo) 2010-07-26 21:04:58 this can be closed - not reproduce-able at this point. By: Paul Belanger (pabelanger) 2010-07-26 21:44:21 Closing, reporter was unable to reproduce. |