> On Feb. 5, 2020, 3:22 a.m., Qian Zhang wrote: > > It seems this test will take 10+ seconds: > > ``` > > [ RUN ] DockerFetcherPluginTest.INTERNET_CURL_GeneratedRepositoryScope > > sh: hadoop: command not found > > I0205 02:02:23.218667 5865 fetcher.cpp:68] Skipping URI fetcher plugin > > 'hadoop' as it could not be created: Failed to create HDFS client: Hadoop > > client is not available, exit status: 32512 > > W0205 02:02:23.721199 5871 docker.cpp:1175] Missing 'realm', 'service' or > > 'scope' in header WWW-Authenticate: Bearer > > realm="https://nvcr.io/proxy_auth?scope=repository:nvidia/k8s/cuda-sample:pull,push" > > [ OK ] DockerFetcherPluginTest.INTERNET_CURL_GeneratedRepositoryScope > > (12176 ms) > > ``` > > > > It is a bit too long, is there a way to accelerate it?
The test spends most of the time uploading the 100Mb of image. Unfortunatley, there seems to be no smaller image on nvcr.io, and I 'm not aware of other repositories suitable for such test. Ideally, we should have a mock repository for this (left a TODO in this patch). - Andrei ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/72080/#review219503 ----------------------------------------------------------- On Feb. 5, 2020, 2:07 p.m., Andrei Sekretenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/72080/ > ----------------------------------------------------------- > > (Updated Feb. 5, 2020, 2:07 p.m.) > > > Review request for mesos, Andrei Budnik and Qian Zhang. > > > Bugs: MESOS-10092 > https://issues.apache.org/jira/browse/MESOS-10092 > > > Repository: mesos > > > Description > ------- > > Added test for fetch from repositories not providing scope/service. > > > Diffs > ----- > > src/tests/uri_fetcher_tests.cpp 92f80626f4de8a2cc29760366ca568f9d794f472 > > > Diff: https://reviews.apache.org/r/72080/diff/3/ > > > Testing > ------- > > > Thanks, > > Andrei Sekretenko > >
