[asterisk-bugs] [JIRA] (ASTERISK-29469) Transfer target remains on hold after remote attended transfer

[Joshua C. Colp (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-29469?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Work on ASTERISK-29469 started by Ben Smithurst.

> Transfer target remains on hold after remote attended transfer
> --------------------------------------------------------------
>
>                 Key: ASTERISK-29469
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-29469
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Resources/res_pjsip_refer
>    Affects Versions: 15.7.2, 18.4.0
>            Reporter: Ben Smithurst
>            Assignee: Ben Smithurst
>              Labels: patch
>         Attachments: transfer-fix.diff
>
>
> We had a problem where a remote attended transfer would leave the transfer destination on hold, with that person only hearing hold music, but the original caller hearing the transfer target.
> We run an Asterisk cluster and customers do not necessarily always receive and place calls via the same server.
> So for example:
> * A calls B (via Asterisk server X)
> * B puts A on hold, and places a new call to the transfer target C (via Asterisk server Y)
> * B speaks to C and then initiates transfer
> * B's phone puts C on hold via server Y and then immediately sends a REFER to Asterisk server X which then handles it via an 'external_replaces' hook, and routes an INVITE-with-Replaces to Asterisk server Y
> * Asterisk server Y completes the transfer BUT doesn't take C off hold.
> The attached patch fixes this problem.  It looks like Asterisk already takes calls off hold when there isn't a remote transfer (i.e. external_replaces hook) involved, so I'm hoping this is correct behaviour.  We've had it in production for around 48 hours now and multiple customers have confirmed it fixes their problem and we've had no new problems reported.
> Please note we've confirmed the fix on 15.7.2 only - we reproduced the problem on 18.4.0 but have not reproduced the problem there as we have reverted our production systems to 15.7.2 for the time being for other reasons.



--
This message was sent by Atlassian JIRA
(v6.2#6252)