v0.1.2

This release is recommended and contains various bugfix.

MD5 (rss-0.1.2-bin-release-with-spark3.tgz) = 3bc9bde4773ce9c383c0506a36ede2b1
MD5 (rss-0.1.2-bin-release-with-spark2.tgz) = 2e8b51f8d019441cc50863b74afd1683

What's Changed

• [ISSUE-496][BUG]Ensure all file channels write whole buffer. (#561) by @waitinfuture in #562
• [ISSUE-567] Optimize LifecycleManager.getLatestPartition by @waitinfuture in #571
• [ISSUE-567][FOLLOW-UP] remove entry from latestPartitionLocation in r… by @waitinfuture in #576
• Cherry-pick transport layer refactor from main branch by @waitinfuture in #587
• [ISSUE-588][BUG] Fix memory leak in shuffle read by @waitinfuture in #590
• [ISSUE-591][BUG] Incorrect result when revive and split happen concur… by @waitinfuture in #593
• [IMPROVEMENT] Improve revive or partition split log readability. by @FMX in #603
• [ISSUE-625][BUG] Incorrect result when kill worker while pushMergedData by @waitinfuture in #626
• [ISSUE-281][BUG] Use correct maxDestLength to check if buffer can satisfy compress result By @AngersZhuuuu in #282
• [BUG]Fix slots release when cleanup expired shuffle By @Flyangz in #514
• [ISSUE-413] Fix incorrect totalSpace in disk usage check By @lichaojacobs in #414
• [Issue 253] Fix netty memory leak in TransportResponseHandler By @dxheming in #254
• [BUG] DiskFlusher nextWorkerIndex update incorrectly By @FMX in #295
• [BUG] After trigger pausePushData, new added channel still can push data By @FMX in #376
• fix clean disk failure when the disk is full. fix concurrent issue in PartitionFilesSorter. @FMX

Full Changelog: v0.1.1...v0.1.2