Feature/log hints #23
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
This is the beginning of code that will reduce the amount of searching needed to find a given index with rafter_log:get_entry. It will keep a dictionary of previously seen index locations in the logfile and use that as a cache for rafter_log:find_entry calls that search the log. As it is very expensive to store pointers to all indexes, this hint strategy will attempt to be smart and only store a subset of frequently needed pointers. Basic stastics are kept to keep track of how efficient the implementation is. It is anticipated in the future that search will find the "closest" hint and not just need an exact cache hit. This should greatly reduce seeks as well. It is also possible in the future that hints will be persisted to their individual log section files.
|
@curiousleo Will you please pull this branch and confirm that this fixes your performance issues? I will then close #22 and merge this PR. |
|
@andrewjstone Sure, I'll do that later today! Looks like an awesome improvement! |
|
Yes, those changes do seem to fix the performance issues I experienced. Great! |
|
Sweet! Tracking this issue down was fun. Performance is fun. Thanks for the On Wed, Feb 12, 2014 at 9:51 AM, Leonhard Markert
|
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Outdated suggestions cannot be applied.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.
Keep track of pointers(hints) to some hot entries in order to minimize disk seeks and data reads. Use a cache eviction policy of removing every tenth hint so we can somewhat uniformly and minimally increase disk seeks with usage in order to maintain a fixed size cache. In other words, when we reach a maximum number of hints we delete 10% of them.
Fixes #22