| Summary: | ASTERISK-00675: [workaround] Using switch => with IAX2 causes block | ||
| Reporter: | bdolljr (bdolljr) | Labels: | |
| Date Opened: | 2003-12-17 11:56:03.000-0600 | Date Closed: | 2004-09-25 02:57:02 |
| Priority: | Blocker | Regression? | No |
| Status: | Closed/Complete | Components: | Core/General |
| Versions: | Frequency of Occurrence | ||
| Related Issues: | |||
| Environment: | Attachments: | ||
| Description: | In this configuration: SIP -> *(1) -> IAX2 -> *(2) -> IAX2 (i.e. digium) When *(1) uses the switch => statement to resolve the exten for an IAX2 destination (i.e digium) on *(2) the asterisk server will block. No additional calls will be processed. Martin (digium) has observed this behavior on our system. | ||
| Comments: | By: zoa (zoa) 2004-01-10 20:20:38.000-0600 any updates on this baby ? could somebody test this ? By: Brian West (bkw918) 2004-01-10 21:54:44.000-0600 Martin is aware of this one but I think they are using a switch in a circular setup which causes this exact situation. By: jrollyson (jrollyson) 2004-01-14 02:40:46.000-0600 This may be a protocol limitation right now, not sure this can be fixed without major changes, but it definantly needs to be fixed. There are ways to engineer the dialplan around this - ie using pattern matching instead of switch on the first hop. By: Brian West (bkw918) 2004-02-01 12:06:00.000-0600 I can't recreate this. I'm going SIP => * => IAX2(switch) => * => IAX2(SWITCH) => * @ Digum. No lock .. it works. By: Brian West (bkw918) 2004-02-01 12:31:04.000-0600 woop just got it to freak out. Very strange its not a deadlock in the normal sense of the word. By: Brian West (bkw918) 2004-02-01 19:31:11.000-0600 Fix added to CVS please verify it corrects this problem. bkw | ||