SPARK-1667 re-fetch fails occasionally #604
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Modified DiskBlockManager to create sub directories when they are absent
sarutak
changed the title
sarutak
changed the title
|
Can one of the admins verify this patch? |
Contributor
There was a problem hiding this comment.
Please dont catch Throwable.
Contributor
|
Can you add a testcase to verify this ? I suspect we have also observed in the past too - but I never got around to nailing it down : thanks for the PR ! |
Member
Author
|
Thank you for your comment, mridulm! I met the bug when I remove the block files in /tmp/spark-local-*/ on the local filesystem of an Executor. |
Contributor
|
Did you mean to close this PR? |
bzhaoopenstack
pushed a commit
to bzhaoopenstack/spark
that referenced
this pull request
Sep 11, 2019
Use old bazel for kind
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I think, when an Executor which has block(s) to be fetched is lost, fetch from the Executor will be fail and re-fetch from another Executor will occur.
When an Executor process is dead, re-fetch will be done successfully, but an Executor process is alive but can't read its own block(s), re-fetch wouldn't occur.
I think the behavior is not expected and tried to modify.
I pull-request my first modification and how does this look?