[KYLIN-3997] Add a health check job of Kylin

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: v3.0.0-alpha2
• Assignee: Shaohui Liu
• Reporter: Shaohui Liu
• Created: 08/May/19 04:50
• Updated: 21/Jan/20 07:43
• Resolved: 09/Jun/19 13:49

Description

Kylin has many inner meta data and outer dependencies. There may be inconsistent for bugs or failures. It's better to have a a health check job to find these inconsistent issues in advance。

The inconsistent issues we found in our clusters are followings

• the cubeid data not exist for cube merging
• hbase table not exist or online for a segment
• there are holes in cube segments(The build of some days failed, but user not found it)
• Too many segment(hbase tables)
• metadata of stale segment  left in cube
• Some cubes have not be updated/built for a long time
• Some  important parameters are no set in cube desc
• ...

 Suggestions are welcomed, thanks~

Issue Links

• links to: GitHub Pull Request #669