| Summary: | ASTERISK-24755: Asterisk sends unexpected early BYE to transferrer during attended transfer when using a Stasis bridge | ||
| Reporter: | John Bigelow (jbigelow) | Labels: | |
| Date Opened: | 2015-02-03 12:43:05.000-0600 | Date Closed: | 2015-03-10 11:06:13 |
| Priority: | Major | Regression? | |
| Status: | Closed/Complete | Components: | Resources/res_pjsip Resources/res_stasis |
| Versions: | SVN 13.2.0 | Frequency of Occurrence | Frequent |
| Related Issues: | |||
| Environment: | Asterisk branch 13 r431492, PJSIP, Stasis, ARI | Attachments: | ( 0) full.txt |
| Description: | Asterisk sends an unexpected early BYE to transferrer during an attended transfer when using a Stasis bridge. After the transferrer completes the transfer, Asterisk sends a BYE to the transferrer for the first dialogue before a 200 OK sipfrag NOTIFY is sent to the transferrer.
The expected behavior is for Asterisk to send the 200 OK sipfrag NOTIFY and to only send a BYE if it doesn't receive a BYE from the endpoint after some amount of time. When using pjsua.py as the endpoint, {{CallCallback.on_transfer_status()}} is never called back after receiving the BYE from Asterisk. The scenario for the attached log (from a test suite test currently in development) can be found here: https://wiki.asterisk.org/wiki/pages/viewpage.action?pageId=30279826#StasisStart/StasisEndTestplan-Attended:Non-StasisApplicationtoStasisBridge | ||
| Comments: | By: John Bigelow (jbigelow) 2015-03-05 23:12:26.157-0600 The patch up for review (r1) appears to resolve the issue. Using the original test suite test where I first ran into this issue: * Patched: 5/5 executions passed * Unpatched: 1/5 executions passed | ||