apache/pinot

Inconsistent segment file extension for TarGzPush and MetadataPush

Open

#10,375 opened on 2023年3月3日

GitHub で見る
 (6 comments) (0 reactions) (1 assignee)Java (1,234 forks)batch import
good first issue

Repository metrics

Stars
 (4,937 stars)
PR merge metrics
 (平均マージ 6d 7h) (30d で 186 merged PRs)

説明

Our SegmentGenerationJobRunner adds tar.gz extension to the segment file name.

  • TarGzPush: the controller side segment upload logic removes tar.gzextension when copying data to deep storage. The segment file will be stored as fileName.
  • MetadataPush: we use pinotFS to copy file from local to dst. Segment file will be stored as fileName.tar.gz.

It's going to be good if we can make the name convension the same. But, this will change the file name convension to the deepstorage. We should at least consider to add the config to remove the extension to have the clear migration path. (e.g. some people can start the cluster using TarGzPush and want to migrate to MetadataPush and keep the file names the same convention in the deep storage)

コントリビューターガイド