Dale, one thing to note on Hoa's patch: We removed the suspending of publishing during the list compare (used to be the full list reload) because it appeared it was no longer necessary to pause events while the list was being compared as very little change actually occurs to the list stored in memory (it hasn't broke yet or had any inaccuracies). Changes to the list are generally very small so they are updated very quickly so an incomplete NOTIFY (per the comments in the old code) seems very unlikely. If you feel this needs to be put back in you will need to address the issue of dead BLF lights for a period of between 10 and 30 seconds on larger systems. 10 to 30 seconds is how long it takes for the RLS service to compare the XML and the list stored in memory. ________________________________________ From: [email protected] [[email protected]] on behalf of Douglas Hubler [[email protected]] Sent: Friday, September 03, 2010 10:22 PM To: sipXecs developer discussions Subject: Re: [sipx-dev] RLS Problems (4.2.1)
On Fri, Sep 3, 2010 at 5:46 PM, Worley, Dale R (Dale) <[email protected]> wrote: > There's a big patch to 4.2 coming over the wall. It includes: > > Fixing this problem > Hoa Nguyen's work on incrementally updating the RLS's data structures > A much more careful handling of delays during the updating process so that > changes are processed faster but also avoid any lockups due to internal > queues filling. > Some bug fixes and minor improvements in related code. What does this mean? Are you attaching a patch somewhere or sending a file to Jean directly? Is this patch the by-product of a svnmerge (hope not) or will we get separate patches for each fix (hope yes)? _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev/ _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev/
