I’m not sure that is what happened on this particular patch. Please note that Jenkins and CH test pool gave same negative result on patch set 2. Meaning both caught an error.
The abnormality here is there are two Jenkins runs on patch set 1. First run failed due to nvme-cli being different version than CH test pool. It happened on quite a few patches, so today morning Karol re-ran the patches which failed due to this particular issue. Seems like new score from patch set 1 overwrote that of patch set 2 (probably due to manual retrigger of tests).
PS. Nvme-cli issue is reported here https://github.com/spdk/spdk/issues/366
I had a typo in a patch last night that failed CH test pool due to a correct scan-build catch however it passed on Jenkins. Can one of you please take a look and see why it passed on Jenkins?
https://review.gerrithub.io/c/spdk/spdk/+/419446 patch set 2 (3 is the fix)