paimon
paimon copied to clipboard
[Improment] The issue of indexing large volumes of data tables.
Search before asking
- [X] I searched in the issues and found nothing similar.
Motivation
Now we have seen significant improvements in query performance through index creation. However, we have encountered several issues during the indexing process.
- Some tasks before sorting and writing to Paimon are getting stuck waiting for data for a long time.
- Due to data skewness, some indexing tasks take a long time to execute.
- The flink batch Job of building index will global failover after taskManager oom
- FilesTable can not query the paritions which has built index
Solution
No response
Anything else?
No response
Are you willing to submit a PR?
- [ ] I'm willing to submit a PR!
issue 1,2 has been fixed related #3081 #2749
issue 3 need a remote suffle service for task failover when using flink