I think this is related to the following review board link:
https://gerrit.asterisk.org/#/c/3320/ When using ODBC storage with VM there is a backlog in processing MWI notify messages. This patch has helped the issue, but not totally resolved it. There is debate on the review board if this is needed by from my experience it is! When using ODBC storage for VM are only registered devices checked for new messages or all peers? Does Asterisk try to send alerts to every peer that has voicemail even if its not registered? ________________________________ From: [email protected] <[email protected]> on behalf of Ross Beer <[email protected]> Sent: 01 August 2016 11:57 To: Asterisk Developers Mailing List Subject: Re: [asterisk-dev] Task Processor Issue ASTERISK-26145 Hi, Just has the same issue on a different test box: pjsip/distributor-000000cc 68046 0 5 450 500 pjsip/distributor-000000cd 68846 0 5 450 500 pjsip/distributor-000000ce 79573 533 12 450 500 pjsip/distributor-000000cf 65862 0 9 450 500 This happens daily on multiple test servers. Can anyone help? Ross ________________________________ From: [email protected] <[email protected]> on behalf of Ross Beer <[email protected]> Sent: 01 August 2016 11:37 To: Asterisk Developers Mailing List Subject: [asterisk-dev] Task Processor Issue ASTERISK-26145 Hi All, I am still experiencing a task processor issue which stops Asterisk 13 processing packets. The task processor shows the following: pjsip/distributor-000000b1 64041 0 6 450 500 pjsip/distributor-000000b2 56970 0 8 450 500 pjsip/distributor-000000b3 59429 0 6 450 500 pjsip/distributor-000000b4 68142 0 8 450 500 pjsip/distributor-000000b5 50941 0 6 450 500 pjsip/distributor-000000b6 56458 0 4 450 500 pjsip/distributor-000000b7 62735 0 5 450 500 pjsip/distributor-000000b8 62856 0 6 450 500 pjsip/distributor-000000b9 68292 0 5 450 500 pjsip/distributor-000000ba 71961 0 4 450 500 pjsip/distributor-000000bb 54810 0 9 450 500 pjsip/distributor-000000bc 60165 0 5 450 500 pjsip/distributor-000000bd 72316 0 10 450 500 pjsip/distributor-000000be 58807 0 8 450 500 pjsip/distributor-000000bf 59188 0 6 450 500 pjsip/distributor-000000c0 66673 0 7 450 500 pjsip/distributor-000000c1 54221 0 5 450 500 pjsip/distributor-000000c2 61391 0 5 450 500 pjsip/distributor-000000c3 78313 527 6 450 500 pjsip/distributor-000000c4 70249 0 4 450 500 pjsip/distributor-000000c5 82573 0 15 450 500 pjsip/distributor-000000c6 63342 0 11 450 500 pjsip/distributor-000000c7 56292 0 10 450 500 pjsip/distributor-000000c8 52082 0 5 450 500 pjsip/distributor-000000c9 68204 0 9 450 500 pjsip/distributor-000000ca 62369 0 4 450 500 pjsip/distributor-000000cb 56135 0 7 450 500 pjsip/distributor-000000cc 62608 0 8 450 500 pjsip/distributor-000000cd 58685 0 5 450 500 pjsip/distributor-000000ce 54513 0 8 450 500 pjsip/distributor-000000cf 56525 0 6 450 500 Processing appears to be fairly evenly distributed across the threads, however, one thread appears to get flooded. I though this could be related to voicemail and MWI https://issues.asterisk.org/jira/browse/ASTERISK-26229 but I am currently testing the patch and it's not showing a queue on the new taskprocessor for app_voicemail. What features use the 'pjsip/distributor' and how can I identify what is causing the backlog? One thing I did notice in the latest back trace, was a call pickup was being done a the time of the issue. Is there a lock that isn't released here? Kind regards, Ross
-- _____________________________________________________________________ -- 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
