My working theory is: - something has caused us to use dramatically more of Docker's "virtual" space (not sure this is the right term) - PVR tests still pass or fail in isolation as one would expect (and presumably we're still down to 127 failures, as we were at the start of this PR) - however, once you've run a certain total number of tests, your local Docker space will be full and everything will fail from that point on
Debugging this doesn't _have_ to hold up this PR, but I don't think there's a huge rush to get it in either, so I can spend some more time on it. I've just resized my local Docker image disk space to 256GB (from 64GB) and am running again to see if it gets further [ Full content available at: https://github.com/apache/beam/pull/6328 ] This message was relayed via gitbox.apache.org for [email protected]
