that would make a lot more sense to me +1
Ed On Jul 13, 2018, at 8:44 AM, Dave Barach (dbarach) <dbar...@cisco.com<mailto:dbar...@cisco.com>> wrote: Makes sense, but it’s a pain when the wheels fall off. Perhaps we should make the virl job non-voting as soon as Maciek thinks we have all P0 test gaps covered? D. From: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> <vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>> On Behalf Of Ed Kern via Lists.Fd.Io Sent: Friday, July 13, 2018 10:11 AM To: Vratko Polak -X (vrpolak - PANTHEON TECHNOLOGIES at Cisco) <vrpo...@cisco.com<mailto:vrpo...@cisco.com>> Cc: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> Subject: Re: [vpp-dev] Revert of gerrit 13408 / "vppinfra: AVX2 blend" in progress.. On Jul 13, 2018, at 5:37 AM, Vratko Polak -X (vrpolak - PANTHEON TECHNOLOGIES at Cisco) via Lists.Fd.Io <vrpolak=cisco....@lists.fd.io<mailto:vrpolak=cisco....@lists.fd.io>> wrote: > Not sure what exactly happened It went like this: 0. Patch set 2 is uploaded. 1. Both clang and virl jobs (among others) are triggered. 2. Both jobs fail (virl on the usual NFS mount symptom). 3. Virl job has naginator, so its result is reported as NOT_BUILT. 4. Clang is reported as FAILURE, so vote is -1. 5. Naginator triggers another run of virl job (only). 6. This time virl job is SUCCESS. 7. Jenkins sees no falure in all 1 job from latest trigger, so votes +1. This is more or less accurate. I am not sure we can explain Jenkins that results from previous trigger rounds still aply (if not superseded). new incoming vote is an overwrite thats just what it does. I recommend to disable naginator on the virl job and rely on methods which trigger all verify jobs. There are no automatic functions to re-trigger all jobs. So your suggesting here is return to a ton of manual rechecks for multiple different possible triggers. The above behavior is not new. It is just rare. If A job fails twice faster than B job fails and then passes you will end up in this state. In this case the clang job is especially prone to this because it is a short duration job that when it fails it does so quickly.. (2,5,6 minutes) So it can actually fail twice (properly) within 10 minutes. The virl job can take 30 minutes (because it does a build first) before it gets to the point where we see intermittent failures. So your looking at @80 minutes to get a proper vote from the virl job (intermittent failure to success). I can remove the retry on virl jobs if that is what the committers would like me to do. But with the virl jobs on the road to deprecation, and hopefully some traction fixing the LF/vexxhost network issues causing so many hudson/jnlp errors, I wouldn’t be voting for this path. Ed Vratko. From: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> <vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>> On Behalf Of Damjan Marion via Lists.Fd.Io Sent: Friday, 2018-July-13 10:59 To: Marco Varlese <mvarl...@suse.de<mailto:mvarl...@suse.de>> Cc: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> Subject: Re: [vpp-dev] Revert of gerrit 13408 / "vppinfra: AVX2 blend" in progress.. Version 1 of the same patch failed properly, for the same error message, then I removed permute inline functions and left blend as i missed that same error happens also for blend. Not sure what exactly happened, but it doesn't look like it is not configured for voting as it worked properly for PatchSet 1 ... Reason why this is failing on clang is that clang guys use __builtin_shufflevector for multiple intel intrinsics, and that builtin insist on immediate value for some parameters, and it refuses to use inline function argument, even if that argument is constant. I will need to redo this with __asm__ ().... -- Damjan On 13 Jul 2018, at 09:49, Marco Varlese <mvarl...@suse.de<mailto:mvarl...@suse.de>> wrote: Hi Dave & allm Sorry about that,I actually merged the patch. However, it was indeed Verified+1 :( I have gone now to see the actual CLANG build failure and wondering: is it possible that job is not configured as voting/gating job in Jenkins so the verification process succeeds anyway? - Marco On Thu, 2018-07-12 at 22:49 +0000, Dave Barach via Lists.Fd.Io wrote: Revert complete... HTH... Dave From: Dave Barach (dbarach) Sent: Thursday, July 12, 2018 5:01 PM To: Damjan Marion (damarion) <damar...@cisco.com<mailto:damar...@cisco.com>>; Florin Coras (fcoras) <fco...@cisco.com<mailto:fco...@cisco.com>>; 'Marco Varlese' <mvarl...@suse.de<mailto:mvarl...@suse.de>>; Ed Kern (ejk) <e...@cisco.com<mailto:e...@cisco.com>> Cc: vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io> Subject: Revert of gerrit 13408 / "vppinfra: AVX2 blend" in progress.. Importance: High Folks, Unfortunately, the AVX2 blend patch causes 100% clang validation failures on unrelated patches. The clang validation job actually failed on the original patch, but somehow fd.io<http://fd.io/> JJB voted +1 anyhow. See https://gerrit.fd.io/r/#/c/13457 – revert, https://gerrit.fd.io/r/#/c/13408 - original patch. Thanks... Dave -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#9830): https://lists.fd.io/g/vpp-dev/message/9830 Mute This Topic: https://lists.fd.io/mt/23297899/675056 Group Owner: vpp-dev+ow...@lists.fd.io<mailto:vpp-dev+ow...@lists.fd.io> Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [mvarl...@suse.de<mailto:mvarl...@suse.de>] -=-=-=-=-=-=-=-=-=-=-=- -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#9837): https://lists.fd.io/g/vpp-dev/message/9837 Mute This Topic: https://lists.fd.io/mt/23297899/675649 Group Owner: vpp-dev+ow...@lists.fd.io<mailto:vpp-dev+ow...@lists.fd.io> Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [e...@cisco.com<mailto:e...@cisco.com>] -=-=-=-=-=-=-=-=-=-=-=-
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#9842): https://lists.fd.io/g/vpp-dev/message/9842 Mute This Topic: https://lists.fd.io/mt/23297899/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
