> On April 26, 2014, 6:10 a.m., Joshua Colp wrote: > > After talking on Twitter I believe I understand what is going on here. > > > > When you do a "sip reload" any realtime peers in chan_sip are pruned and > > removed. This is probably causing the device state to change to an > > incorrect value which app_queue is then using. The change probably works > > around this by re-checking using any channels that exist. > > Matt Jordan wrote: > If that's the case, this solution still isn't the correct solution to the > problem. The fact that the device state isn't being persisted correctly > across the reload would point to a problem in chan_sip. > > While I appreciate the attempt to fix this problem, adding complexity to > an already complex module that doesn't address the root cause of this problem > is not an appropriate solution. If you'd like to dig into the reload code in > chan_sip to try and find why the realtime peers are getting the wrong device > state, that'd be appreciated - I think everyone would be happy to discuss the > situation more in #asterisk-dev or on the asterisk-dev mailing list. If > you're interested in providing a patch that solves the problem where it is > occurring, I'll keep this review open for that patch - otherwise, barring any > objections, I'll close this out in a few days.
So I have to wonder - is this problem solved by the approach on https://reviewboard.asterisk.org/r/3508/? - Matt ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviewboard.asterisk.org/r/3409/#review11759 ----------------------------------------------------------- On April 2, 2014, 7:48 a.m., Shlomi Gutman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviewboard.asterisk.org/r/3409/ > ----------------------------------------------------------- > > (Updated April 2, 2014, 7:48 a.m.) > > > Review request for Asterisk Developers. > > > Bugs: ASTERISK-16115 > https://issues.asterisk.org/jira/browse/ASTERISK-16115 > > > Repository: Asterisk > > > Description > ------- > > In some cases when member in talk (IN_USE) and you run "sip reload" or peer > has connectivity problems (reachable->unreachable/lagged->reachable) the > status of peer is set to NOT_INUSE, while he is still talking. > The patch adds function that would check if member has any active channel. > And would consider him IN_USE if does and if not, would do nothing, as you > can't detect what is the state of member if he has no channel. > > Originally patch was used on 1.8.23.0 (can apply it to reviewboard as well) > and 11.6-cert1. As well as in my case all peers, queues, memebrs are > realtime. But as David Brillert reported in bug, he is not using realtime > peers, so it's used for all peers. > > The patch works, the question is if it's the best solution and it would not > introduce any regressions. > > > Diffs > ----- > > /branches/11/apps/app_queue.c 411575 > > Diff: https://reviewboard.asterisk.org/r/3409/diff/ > > > Testing > ------- > > Used in production on 1.8 and 11 and customers having problems approved that > problem was resolved. > > > Thanks, > > Shlomi Gutman > >
-- _____________________________________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-dev mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev