Yes, that is my suggest, merge the revert in https://git.opendaylight.org/gerrit/#/c/53953 <https://git.opendaylight.org/gerrit/#/c/53953> as I do not think we have time to investigate and fix the perf issues.
> On Mar 28, 2017, at 11:26 AM, Abhijit Kumbhare <[email protected]> wrote: > > So Luis - you are suggesting not to merge Tomas' patch. You are also > suggesting that I should merge the revert by Jozef: > https://git.opendaylight.org/gerrit/#/c/53953/ > <https://git.opendaylight.org/gerrit/#/c/53953/> . Can you please confirm? > > On Tue, Mar 28, 2017 at 10:52 AM, Luis Gomez <[email protected] > <mailto:[email protected]>> wrote: > Yes, I think revert is more safe now, we go back to bug 6917 but that was > failing since Boron release. Tomas, I tried with your patch and I still see > device connection issues and stats not being collected for 4K flows in 1 > switch and 20K flows in 8 switches. > > BR/Luis > > >> On Mar 28, 2017, at 9:39 AM, Tomáš Slušný <[email protected] >> <mailto:[email protected]>> wrote: >> >> Luis - Great, thank you. >> Abhijit - In worst case this can reintroduce bug 6917, but as I said, I ran >> test-openflowplugin-core on it, and as part of that tests it is checking for >> bug 6917, and this test case passed, so that looks okay. Other than that, it >> will not break anything else, as that patch I created is basically partial >> revert to how DeviceFlowRegistry worked before patch that resolved 6917 was >> merged. >> >> Tomas >> Od: Abhijit Kumbhare <[email protected] <mailto:[email protected]>> >> Odoslané: 28. marca 2017 18:29 >> Komu: Luis Gomez >> Kópia: Tomáš Slušný; Jozef Bacigál; openflowplugin-dev >> Predmet: Re: [openflowplugin-dev] Last patch in boron introduces perf >> regression >> >> Just wanted to ask - how safe is this fix Tomas? >> >> On Tue, Mar 28, 2017 at 8:36 AM, Luis Gomez <[email protected] >> <mailto:[email protected]>> wrote: >> I cherry-picked your patch, as we are interested in boron now: >> >> https://git.opendaylight.org/gerrit/#/c/53989 >> <https://git.opendaylight.org/gerrit/#/c/53989> >> >> I will test it manually once it verifies. >> >> BR/Luis >> >> >>> On Mar 28, 2017, at 5:56 AM, Tomáš Slušný <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> I tried to fix this performance issue in this patch: >>> https://git.opendaylight.org/gerrit/#/c/53972/ >>> <https://git.opendaylight.org/gerrit/#/c/53972/>. I also ran >>> test-openflowplugin-core and it seems that even with this fix bug 6917 test >>> cases are still passing on Jenkins, so Luis can you check if that patch >>> solves that performance regression or not? >>> >>> Tomas >>> Od: Jozef Bacigál >>> Odoslané: 28. marca 2017 9:33 >>> Komu: Jozef Bacigál; Abhijit Kumbhare; Luis Gomez; openflowplugin-dev; >>> Tomáš Slušný >>> Predmet: RE: [openflowplugin-dev] Last patch in boron introduces perf >>> regression >>> >>> Anyway Luis, >>> >>> Maybe I can’t read the proper the tests, but how do you see there is >>> regression ? :) >>> >>> Jozef >>> >>> From: Jozef Bacigál [mailto:[email protected] >>> <mailto:[email protected]>] >>> Sent: Tuesday, March 28, 2017 8:30 AM >>> To: Abhijit Kumbhare <[email protected] >>> <mailto:[email protected]>>; Luis Gomez <[email protected] >>> <mailto:[email protected]>>; openflowplugin-dev >>> <[email protected] >>> <mailto:[email protected]>>; Tomáš Slušný >>> <[email protected] <mailto:[email protected]>> >>> Subject: Re: [openflowplugin-dev] Last patch in boron introduces perf >>> regression >>> >>> Abhijit so in the master branch no problem we got plenty of time, but you >>> say for the SR3 better with race condition than regression ? >>> >>> P.S.: Right now all merges are blocked assume so we can only prepare revert. >>> >>> Here is the revert for boron: >>> >>> https://git.opendaylight.org/gerrit/#/c/53953/ >>> <https://git.opendaylight.org/gerrit/#/c/53953/> >>> >>> Jozef. >>> >>> >>> From: Abhijit Kumbhare [mailto:[email protected] >>> <mailto:[email protected]>] >>> Sent: Tuesday, March 28, 2017 7:21 AM >>> To: Luis Gomez <[email protected] <mailto:[email protected]>>; >>> openflowplugin-dev <[email protected] >>> <mailto:[email protected]>> >>> Subject: Re: [openflowplugin-dev] Last patch in boron introduces perf >>> regression >>> >>> Sounds good. Jozef/Tomas can you do it? I assume the race condition that >>> the patch fixed was pre-existing. >>> >>> On Mon, Mar 27, 2017 at 9:12 PM Luis Gomez <[email protected] >>> <mailto:[email protected]>> wrote: >>> Hi all, >>> >>> After looking these boron jobs: >>> >>> https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-1node-periodic-scale-stats-collection-daily-only-boron/ >>> >>> <https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-1node-periodic-scale-stats-collection-daily-only-boron/> >>> https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-3node-periodic-bulkomatic-clustering-daily-only-boron/ >>> >>> <https://jenkins.opendaylight.org/releng/view/openflowplugin/job/openflowplugin-csit-3node-periodic-bulkomatic-clustering-daily-only-boron/> >>> >>> and testing SR3 manually, I conclude last patch we merged to fix stats >>> issues: >>> >>> https://git.opendaylight.org/gerrit/#/c/53545 >>> <https://git.opendaylight.org/gerrit/#/c/53545> >>> >>> Introduces *important* perf regression, so I really recommend to revert it >>> before we release SR3. >>> >>> Same thing I recommend for master but there we have more time to revert/fix: >>> >>> https://git.opendaylight.org/gerrit/#/c/52237 >>> <https://git.opendaylight.org/gerrit/#/c/52237> >>> >>> BR/Luis >>> >>> _______________________________________________ >>> openflowplugin-dev mailing list >>> [email protected] >>> <mailto:[email protected]> >>> https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev >>> <https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev> >> >> > >
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
