The intake code ensures that all dataset (even server.archiveonly datasets) will have a server.benchmark value representing the core benchmark used in the run. The visualization code uses this value to determine compatibility.
However on the staging server, we have pre-existing datasets without the server.benchmark metadata, and the "internal server error" failure mode is unpleasant.
To be a bit more friendly, this adds a wrapper that will first check the server.benchmark metadata directly, but failing that will check for the normal Pbench Agent dataset.metalog.pbench.script metadata so that older pbench-uperf runs will be recognized. Failing that, the wrapper falls back to "unknown" so that we'll never have to deal with a None value.
PBENCH-1229
The intake code ensures that all dataset (even
server.archiveonlydatasets) will have aserver.benchmarkvalue representing the core benchmark used in the run. The visualization code uses this value to determine compatibility.However on the staging server, we have pre-existing datasets without the
server.benchmarkmetadata, and the "internal server error" failure mode is unpleasant.To be a bit more friendly, this adds a wrapper that will first check the
server.benchmarkmetadata directly, but failing that will check for the normal Pbench Agentdataset.metalog.pbench.scriptmetadata so that olderpbench-uperfruns will be recognized. Failing that, the wrapper falls back to "unknown" so that we'll never have to deal with aNonevalue.