I would like these generic tests suites to have specific purpose, not scattered around the codebase.
Cool. Note that "generic" means that the purpose might not be "specific".
Does it…
Not sure there's an actual difference for us between a "blob" and an "object" from the storage viewpoint. Why to move it there?
Anyway, it's intended to be used everywhere it can be plugged in.…
They should be executed in settings that are representative of how they will be used; we can't arbitrarily change parameters in a production deployment to match a specific workload.
This…
Speaking of benchmarks -- sequential benchmarks are affected by the batch delay and batch size. Have you made batch_size=1 for bbolt? Parallel benchmarks need to be executed in similar…
Results from basic benchmarks:
$ go test -benchmem -run=^$ -benchtime 1m -cpu 1 -timeout 1h -bench ^BenchmarkWritecacheSeq$ git.frostfs.info/TrueCloudLab/frostfs-node/pkg/local_object_stora…@ale64bit I thought Generic tests are exactly like this (and more comprehensive) -- it seems not.
In this particular case, the test caught it because it deletes twice the same key, which…
Are these errors caught by tests?
Not currently, AFAIK. I discovered it with a test for the lite cache: TrueCloudLab/frostfs-node#654/files…