Liangjun He
Liangjun He
> Please fix the conflicts. And how do you plan to reduce the pressure on NN when fetching the accessTime? Or is this not a big problem? > > Thanks....
@Hexiaoqiao sir. Thank you for your review, it caused by enable http UI.
@Hexiaoqiao OK Sir, I have already fixed and committed it.
@xinglin @ayushtkn @tomscut Thank you for your review, and I have a new commit for the above issues.
@slfan1989 @xinglin @Hexiaoqiao Are there any other suggestions for this PR?
@Suave Is Hadoop SDK a better choice for big data scenarios?
@Apache9 sir. Could you take a look? Thanks.
The current implementation sets a tiered storage policy for the bulktoken/family directory, which is cleaned up after bulkloading. Therefore, I haven't figured out how to verify it through UT.
This patch has already been used by our customers in production environments.
@NihalJain Thanks for your review. The Bulkload process consists of two steps: 1. generate hfiles using MR/SPARK and write them to an HDFS cluster. 2. execute 'hbase completebulkload [OPTIONS] '...