Github user NicoK commented on the issue:
https://github.com/apache/flink/pull/3512
let's include the improvements in a separate PR #4146 and drop the feature
additions for a re-work in FLIP-19
---
If your project is set up for it, you can reply to this email and have your
reply app
Github user zentol commented on the issue:
https://github.com/apache/flink/pull/3512
I take it this PR will be subsumed in FLIP-19?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this featur
Github user NicoK commented on the issue:
https://github.com/apache/flink/pull/3512
After investigating a bit further, I noticed that this problem is actually
a bit bigger:
Even in `FileSystemBlobStore`, there is no guarantee that a directory will
not be deleted concurrently (from
Github user NicoK commented on the issue:
https://github.com/apache/flink/pull/3512
Found a race between `BlobCache#deleteAll(JobID)` and
`BlobCache#getURL(BlobKey)` now that the former is actually being used - this
needs to be fixed first before merging:
`BlobCache#deleteAll
Github user NicoK commented on the issue:
https://github.com/apache/flink/pull/3512
* I removed the exposed `BlobService` from the `LibraryCacheManager`
* Also, I developed a new cleanup story that removes blobs only if there
are no tasks referring to the job ID anymore. The extens