Add option for smaller bitcask keys #643
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
The new keydir format has 2 forms, as follows <<Version:7, 0:1, BucketSize:16/integer, Bucket/binary, Key/binary>> <<Version:7, 1:1, TypeSize:16/integer, Type/binary, BucketSize:16/integer, Bucket/binary, Key/binary>> The 8th bit of the binary is the bit that indicates if a bucket type is present, and tells the decoder which form to expect.
|
This mostly looks ok, although it needs my refactor to support bucket types over in #663. |
Modify the version 1 format to support bucket types
|
Merged, your PR, is this ready to merge once the bitcask comments have been addressed? |
|
Yes. |
evanmcc
added a commit
that referenced
this pull request
Sep 20, 2013
Add option for smaller bitcask keys
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Outdated suggestions cannot be applied.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.
The kv bucket/key to bitcask key conversion uses term_to_binary. This results in 13 bytes of overhead per key.
This patch introduces a new key format:
The first two bytes specify the length of the bucket, allowing kv to translate the compact binary back to the bucket/key pair that it needs. In order to read old-format values off of disk, this code allows the insertion of a translation function, so that keys in the keydir are in a consistent format. This saves 11 bytes per entry in the keydir.
When a large percentage of your keys on disk are in the old format, there is a small but noticable slowdown in startup time and fold time. This should go down over time as keys are replaced. If the slowdown is intolerable, adding
{small_keys, false}to the bitcask section of your app.config will revert to the old format. This feature is currently on by default; whether this should be the case is up for debate.This PR depends on a change in bitcask adding support for key transformation functions:
basho/bitcask#103