All - In 3 days I plan to collapse into master the volatile/patch-tracking/5/accepted branch.
If you have concerns/questions please speak now. thanks! donald On Fri, Dec 4, 2015 at 7:36 AM, Donald Sharp <[email protected]> wrote: > Martin - > > Thanks for your testing! > > donald > > On Fri, Dec 4, 2015 at 3:42 AM, Martin Winter < > [email protected]> wrote: > >> Everything below applies to latest proposed-5 branch as on Savannah, >> Git commit 496325d91 (as of Nov 25) >> >> First of all, this took way longer than what I’ve expected. >> The change of the default in BGP caused me to loose basically a week of >> testing… >> >> As some background: I’m using Ixia ANVL for testing and this software has >> it’s own >> IP stack - and sends very specific and controlled packets, i.e. there is >> no such thing >> as TCP retransmission in it (It would detect something like this as a >> failure) and >> it won’t negotiate about timers - it expects them to be as configured in >> the test setup. >> >> I’ve now adapted all (or at least most) of my tests to deal with this >> change (i.e. for the BGP >> timers, I’m going to configure them now in the test script instead of >> assuming some value) >> >> Anyway, I’m quite certain that all the new BGP failures are based on the >> change of the timers. >> OSPFv2 and OSPFv3, ISIS IPv4, RIP and RIPNG look ok >> >> On ISIS IPv6, there is one new failure where I do believe that ANVL is >> wrong (ANVL >> checking the wrong LSP) >> >> Basically, I think Proposed-5 branch is good (from my side) to be merged >> into Master. >> >> Regards, >> Martin Winter >> [email protected] >> >> >> Results are updated at >> >> https://drive.google.com/folderview?id=0B8W_T0dxQfwxSHR6UGk4RDFteEk&usp=sharing >> >> >> >> On 25 Nov 2015, at 3:40, Martin Winter wrote: >> >> Latest round of proposed-5 is basically equal. >>> >>> Results are here: >>> >>> https://drive.google.com/folderview?id=0B8W_T0dxQfwxSHR6UGk4RDFteEk&usp=sharing >>> >>> I’ll start with a git bisect on a few of the failures over the next day. >>> Expect more details later this week. >>> >>> Regards, >>> Martin Winter >>> >>> On 23 Nov 2015, at 19:55, Martin Winter wrote: >>> >>> Ok, here is the summary of the changes in test results for the proposed >>>> 5 branch >>>> (as of Nov 17, Git commit 5ab56c73 - so NOT CURRENT) >>>> >>>> BGP IPv4: 4 new failures, 4 new unpredictable (before all >>>> passed) >>>> BGP IPv4 AS4: No change >>>> BGP IPv6: 18 new pass (before failed or unpredictable), 3 >>>> new failures (before pass) >>>> BGP IPv6 AS4: No change >>>> OSPFv2 (IPv4): 1 new pass (before fail), 1 new fail, 1 new >>>> unpredictable (before passed) >>>> OSPFv3 (IPv6): 3 new pass (before unpredictable), 1 new unpredictable >>>> (before pass) >>>> ISIS IPv4: 2 new pass (before fail), 1 new pass (before >>>> unpredictable) >>>> ISIS IPv6: 2 new pass (before fail), 3 new fail (before >>>> pass) >>>> RIP (IPv4): no change >>>> RIPNG (IPV6): 1 new pass (before unpredictable) >>>> >>>> Comparison is against the final accepted-4 branch (which should be same >>>> as current >>>> master - but I had no time yet to actually run the current master) >>>> >>>> I have not looked into details on the changes. I’ll start looking into >>>> them >>>> probably after the next update for the proposed 5 branch >>>> (All my testbeds are currently busy running the current proposed 5 >>>> branch) >>>> >>>> I’ll post updated spreadsheets online for everyone after the current >>>> proposed 5 >>>> branch is done with the testrun (expected by noon GMT on Wed) >>>> >>>> Regards, >>>> Martin Winter >>>> >>>> >>>> On 23 Nov 2015, at 13:58, Martin Winter wrote: >>>> >>>> Just as a heads up: >>>>> >>>>> I should have full RFC compliance results for proposed 5 as of git >>>>> commit 5ab56c73c >>>>> (Nov 17 - when the BSD and Solaris issue still existed) in approx >>>>> another 6 hrs. >>>>> >>>>> Current version is started as well and results should be available in >>>>> approx 36 hrs >>>>> >>>>> (Doesn’t mean that you need to wait - this is just a heads up for your >>>>> planning) >>>>> >>>>> - Martin >>>>> >>>>> >>>>> On 23 Nov 2015, at 9:06, Donald Sharp wrote: >>>>> >>>>> Paul - >>>>>> >>>>>> I don't mind pushing them up, but I wanted to solve the issue that >>>>>> Martin >>>>>> has discovered about the FSM log message first if you don't mind? >>>>>> >>>>>> donald >>>>>> >>>>>> On Mon, Nov 23, 2015 at 5:09 AM, Paul Jakma <[email protected]> wrote: >>>>>> >>>>>> On Mon, 16 Nov 2015, Donald Sharp wrote: >>>>>>> >>>>>>> Currently everything queued up is at: >>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> https://github.com/donaldsharp/quagga/tree/volatile/patch-tracking/5/proposed/ff >>>>>>>> >>>>>>>> >>>>>>> Can you push them to Savannah? >>>>>>> >>>>>>> Savannah doesn't allow non-ff updates of branches, so after a rebase >>>>>>> of a >>>>>>> published volatile branch you need to delete it and recreate >>>>>>> (*never* do >>>>>>> this for master). This is what I do: >>>>>>> >>>>>>> for H in deferred accepted rejected ; do >>>>>>> git push quagga-gnu.org :refs/heads/volatile/patch-tracking/4/${H} >>>>>>> git push quagga-gnu.org >>>>>>> ${H}:refs/heads/volatile/patch-tracking/4/${H} >>>>>>> done >>>>>>> >>>>>>> Where quagga-gnu.org is the remote I have defined: >>>>>>> >>>>>>> $ git remote -v | grep quagga-gnu >>>>>>> quagga-gnu.org ssh://git.sv.gnu.org/srv/git/quagga.git (fetch) >>>>>>> quagga-gnu.org ssh://git.sv.gnu.org/srv/git/quagga.git (push) >>>>>>> >>>>>>> Btw, you havn't carried forward the previous round's deferred branc? >>>>>>> >>>>>>> regards, >>>>>>> -- >>>>>>> Paul Jakma [email protected] @pjakma Key ID: 64A2FF6A >>>>>>> Fortune: >>>>>>> God, I ask for patience -- and I want it right now! >>>>>>> >>>>>>> >
_______________________________________________ Quagga-dev mailing list [email protected] https://lists.quagga.net/mailman/listinfo/quagga-dev
