[ https://issues.apache.org/jira/browse/KNOX-1525?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16650490#comment-16650490 ]
Kevin Risden edited comment on KNOX-1525 at 10/15/18 5:25 PM: -------------------------------------------------------------- h2. Result Interpretation The HBase shell since it is built in Ruby is not the best tool for scanning HBase. Typically the HBase Java API would be used. HBase Rest performs significantly better than HBase shell. HBase Rest with Knox also performs very well with only a slight overhead of <2 seconds for 400 thousand rows. There doesn't seem to be any glaring bottlenecks when using Knox with HBase Rest. was (Author: risdenk): h2. Result Interpretation TODO > HBase "scan" performance evaluation > ----------------------------------- > > Key: KNOX-1525 > URL: https://issues.apache.org/jira/browse/KNOX-1525 > Project: Apache Knox > Issue Type: Task > Reporter: Kevin Risden > Assignee: Kevin Risden > Priority: Major > Fix For: 1.2.0 > > > While looking at WebHDFS performance in KNOX-1221, I decided to look a bit > more into performance for common use cases. HBase performance is another area > that could use some research. > Use "scan limit" to get a comparison of raw return speed from HBase Rest. > This should show how fast results can be streamed through HBase Rest and > Knox. Compare the results to "hbase shell" since this will render the data > directly from HBase through the typical HBase shell. This should give > comparisons for the difference in overhead between HBase, HBase Rest and > HBase Rest with Knox. -- This message was sent by Atlassian JIRA (v7.6.3#76005)