OK.
Well the fix I am doing will completely alter the mechanism for generating NOTIFYs when subs_db_mode is DB only and a couple of other parameters are set. However, I am deliberately leaving the existing behaviour unchanged for other modes. So it sounds like there is a problem here, but not one my work will fix.
Peter
On Wed, 2012-05-02 at 12:15 +0300, Juha Heinanen wrote:
Peter Dunkley writes:
There are lots of CSeq race conditions when subs_db_mode is DB only. I am working on adding separate notifier processes to presence which, amongst other things, will fix this.
peter,
nice to hear. in my test i had not set subs_db_mode, which means Write-Back scheme. also, i don't think that in this test race conditions play any role, because time difference between the two notifies was 19 seconds. looks to me the this bug is of more fundamental nature.
-- juha
sr-dev mailing list sr-dev@lists.sip-router.org http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev