Danil Lipovoy created HADOOP-16901:
--------------------------------------
Summary: HDFS-client: boost ShortCircuit Cache
Key: HADOOP-16901
URL: https://issues.apache.org/jira/browse/HADOOP-16901
Project: Hadoop Common
Issue Type: New Feature
Environment: 4 nodes E5-2698 v4 @ 2.20GHz, 700 Gb Mem.
8 RegionServers (2 by host)
8 tables by 64 regions by 1.88 Gb data in each = 1200 Gb total
Random read in 800 threads via YCSB and a little bit updates (10% of reads)
Reporter: Danil Lipovoy
Attachments: hdfs_cpu.png, hdfs_reads.png
I want to propose how to improve reading performance HDFS-client. The idea:
create few instances SchortCircuit caches instead of one.
The key points:
1. Create array of caches:
{code:java}
private ClientContext(String name, DfsClientConf conf, Configuration config) {
...
shortCircuitCache = new ShortCircuitCache[this.clientShortCircuitNum];
for (int i = 0; i < this.clientShortCircuitNum; i++) {
this.shortCircuitCache[i] = ShortCircuitCache.fromConf(scConf);
}
{code}
2 Then divide blocks by caches:
{code:java}
public ShortCircuitCache getShortCircuitCache(long idx) {
return shortCircuitCache[(int) (idx % clientShortCircuitNum)];
}
{code}
3. And how to call it:
{code:java}
ShortCircuitCache cache =
clientContext.getShortCircuitCache(block.getBlockId());
{code}
The last number of offset evenly distributed from 0 to 9 - thats why all caches
will full approximatly the same.
It is good for performance. Below the attachment, where clientShortCircuitNum =
3. There is load test reading HDFS via HBase. We can see that performance grows
~30%, CPU usage about 15%.
Will try to add the link to PullRequest soon.
Hope it is intresting for somebody.
Ready to explain some unobvious things.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-dev-h...@hadoop.apache.org
