3
61%
main: 61%

Ran 18 Dec 2018 03:50AM UTC

Files 28

Run time 1s

Badge

Embed ▾

Committed 18 Dec 2018 03:44AM UTC coverage: 79.609% (-0.5%) from 80.096%

Job # 1719.3

Build Type

push

travis-ci

Committed by

web-flow

Commit Message

New tools, new option (#645)

- Add a new flatten command, which would flatten out all the LSM tree levels into a single level. It ensures that that level does not exceed its max size limits, as specified in the options.
- Add a new fill command, which would random key-values to Badger. Useful for testing.

## Option updates
- Add a new option to skip force compaction of Level 0 on DB.Close.
- Decrease the NumCompactors default value from 3 to 2.

Commits:

* Add two new commands: flatten and fill.
* Add an option to skip compacting Level 0 for performance reasons.
* Allow more compactors to run for flatten. Performance wise I didn't see any gains in latency while using multiple compactors. And there's a lot of resource consumption if multiple compactors are running concurrently. Could be due to all compactors working on the same level during flatten. In a live setting, compactors could be working on different levels, so the result might not directly apply.
* Decreased the NumCompactors default option to 2. Didn't set it to 1 to avoid a straggler compactor blocking all compactions.
* Add licenses.

Run Details

4685 of 5885 relevant lines covered (79.61%)

62840.09 hits per line

dgraph-io / badger / 1719 / 3
61%
main: 61%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 1719.3

dgraph-io / badger / 1719 / 3 61% main: 61%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 1719.3

dgraph-io / badger / 1719 / 3
61%
main: 61%

README BADGES
x