New S3 object storage implementation / BlobStore #1030
Merged
Conversation
This file contains 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
bryanlb
force-pushed
the
bburkholder/new-object-storage
branch
from
664aa34 to
5a396a8
Compare
bryanlb
force-pushed
the
bburkholder/new-object-storage
branch
2 times, most recently
from
8085869 to
89b7d25
Compare
bryanlb
force-pushed
the
bburkholder/new-object-storage
branch
from
89b7d25 to
04090c5
Compare
bryanlb
changed the title
2 tasks
bryanlb
force-pushed
the
bburkholder/new-object-storage
branch
from
1c5205c to
3d7fae1
Compare
bryanlb
changed the title
kyle-sammons
approved these changes
Aug 26, 2024
2 tasks
zarna1parekh
pushed a commit
to airbnb/kaldb
that referenced
this pull request
Sep 5, 2024
* Initial chunk store implementation * Remove unnecessary object storage code, part 1 * Add listFiles, delete to chunkStore implementation, remove from blobfs * Remove use of blobfsutils copyFromS3 test helper * Remove remaining blobfs implementations * Remove snapshotpath, index type from snapshots * Refactoring, add documentation, add tests --------- Co-authored-by: Bryan Burkholder <[email protected]>
3 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Summary
Refactors the existing object storage implementation. The current
blobfsimplementation is overly abstracted, and as such requires a lot of unnecessary list API calls to S3 (#850). This causes significant performance issues as the number of items in the bucket increases.A clean sheet design was implemented here to reduce the abstractions, and setup the design so that we can take advantage of https://github.com/awslabs/aws-java-nio-spi-for-s3 in a future PR.
Initial performance benchmarks for large S3 buckets (large in this context being ~ 1PB/1.5M items) are up to 40% faster for downloading. Delete operations / chunk expirations have also seen significant speed increases.
This also removes the snapshot
pathandindexTypefields - the former because all data is now accessed via the snapshot ID, and the latter because it is unused/unnecessary.