A testbench test would be much better than a manual repo. I was about to create one before libfastjson and package system work. The v8 engine is totally different. I think we could indeed extend the transaction entry point, thought this out ~50%. But there is still many things to do and time is sparse. If the prep work for the testbench would be done, that would probably speed up the real work.
Rainer Sent from phone, thus brief. Am 20.04.2016 21:23 schrieb "David Lang" <[email protected]>: > We have one good example of a problem that's easy to reproduce. > > http://lists.adiscon.net/pipermail/rsyslog/2016-March/042379.html > > In that thread you said that there were changes with the v8 batch handling > that were needed for things like the global variables to work that were a > problem, but you were busy at the time and were never able to follow up on > the topic. > > David Lang > > > > On Wed, 20 Apr 2016, Rainer Gerhards wrote: > > Date: Wed, 20 Apr 2016 21:01:39 +0200 >> From: Rainer Gerhards <[email protected]> >> Reply-To: rsyslog-users <[email protected]> >> To: rsyslog-users <[email protected]> >> Subject: Re: [rsyslog] rsyslog's batch api documentation inconsistency >> >> That's what is intended, all else is a bug. >> >> If you look at the code, check end transaction Handling. A good start >> would >> be to create a test bench test that exposes the problem. So we have >> something consistent to work on and also guard against future such issues. >> Am 20.04.2016 20:50 schrieb "Brian Knox" <[email protected]>: >> >> I think at least once is the best choice as well. >>> >>> On Wed, Apr 20, 2016 at 2:18 PM Kane Kim <[email protected]> wrote: >>> >>> Yes, we would strongly prefer at-least-once semantics. If everyone agrees >>>> that it's desired behavior for rsyslog I'll try to contribute something >>>> >>> in >>> >>>> that direction. >>>> >>>> On Wed, Apr 20, 2016 at 10:38 AM, David Lang <[email protected]> wrote: >>>> >>>> On Wed, 20 Apr 2016, Kane Kim wrote: >>>>> >>>>> According to documentation ( >>>>> >>>>>> >>>>>> >>>>>> >>>> >>> http://www.rsyslog.com/doc/v8-stable/development/dev_oplugins.html#batching-of-messages >>> >>>> ) >>>>>> rsyslog's behavior with batch api output plugins should be >>>>>> >>>>> at-least-once. >>>> >>>>> Right now it's at-most-once. >>>>>> >>>>>> What's the preferred approach, fix rsyslog to provide at-least-once >>>>>> guarantees or fix documentation so it mentions that rsyslog doesn't >>>>>> provide >>>>>> it? >>>>>> >>>>>> >>>>> I think it should be at least once. It's a lot easier to weed out >>>>> duplicates than to notice that something is missing. >>>>> >>>>> David Lang >>>>> _______________________________________________ >>>>> rsyslog mailing list >>>>> http://lists.adiscon.net/mailman/listinfo/rsyslog >>>>> http://www.rsyslog.com/professional-services/ >>>>> What's up with rsyslog? Follow https://twitter.com/rgerhards >>>>> NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a >>>>> >>>> myriad >>> >>>> of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you >>>>> DON'T LIKE THAT. >>>>> >>>>> _______________________________________________ >>>> rsyslog mailing list >>>> http://lists.adiscon.net/mailman/listinfo/rsyslog >>>> http://www.rsyslog.com/professional-services/ >>>> What's up with rsyslog? Follow https://twitter.com/rgerhards >>>> NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad >>>> of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you >>>> DON'T LIKE THAT. >>>> >>>> _______________________________________________ >>> rsyslog mailing list >>> http://lists.adiscon.net/mailman/listinfo/rsyslog >>> http://www.rsyslog.com/professional-services/ >>> What's up with rsyslog? Follow https://twitter.com/rgerhards >>> NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad >>> of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you >>> DON'T LIKE THAT. >>> >>> _______________________________________________ >> rsyslog mailing list >> http://lists.adiscon.net/mailman/listinfo/rsyslog >> http://www.rsyslog.com/professional-services/ >> What's up with rsyslog? Follow https://twitter.com/rgerhards >> NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad >> of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you >> DON'T LIKE THAT. >> >> _______________________________________________ > rsyslog mailing list > http://lists.adiscon.net/mailman/listinfo/rsyslog > http://www.rsyslog.com/professional-services/ > What's up with rsyslog? Follow https://twitter.com/rgerhards > NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad > of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you > DON'T LIKE THAT. > _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE THAT.
