[jira] [Commented] (HBASE-9775) Client write path perf issues

[Elliott Clark (JIRA)]

    [ 
https://issues.apache.org/jira/browse/HBASE-9775?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13796294#comment-13796294
 ] 

Elliott Clark commented on HBASE-9775:
--------------------------------------

YCSB with 33 nodes fails completely. with:

{code}
com.yahoo.ycsb.DBException: 
org.apache.hadoop.hbase.client.RetriesExhaustedWithDetailsException: Failed 
8608 actions: IOException: 8608 times, 
        at com.yahoo.ycsb.db.HBaseClient.cleanup(HBaseClient.java:113)
        at com.yahoo.ycsb.DBWrapper.cleanup(DBWrapper.java:73)
        at com.yahoo.ycsb.ClientThread.run(Client.java:307)
Caused by: org.apache.hadoop.hbase.client.RetriesExhaustedWithDetailsException: 
Failed 8608 actions: IOException: 8608 times, 
        at 
org.apache.hadoop.hbase.client.AsyncProcess$BatchErrors.makeException(AsyncProcess.java:185)
        at 
org.apache.hadoop.hbase.client.AsyncProcess$BatchErrors.access$500(AsyncProcess.java:169)
        at 
org.apache.hadoop.hbase.client.AsyncProcess.getErrors(AsyncProcess.java:782)
        at 
org.apache.hadoop.hbase.client.HTable.backgroundFlushCommits(HTable.java:934)
        at org.apache.hadoop.hbase.client.HTable.flushCommits(HTable.java:1193)
        at com.yahoo.ycsb.db.HBaseClient.cleanup(HBaseClient.java:108)
        ... 2 more
{code}

running:

{code}
#!/bin/bash
OPERTATIONS=20000000
RECORDS=$(($OPERTATIONS*10))

echo "disable 'usertable'; drop 'usertable'" | hbase shell
echo "create 'usertable',{NAME => 'd', VERSIONS => 1,COMPRESSION => 'lz4', 
BLOCKCACHE => true}" | hbase shell
echo "balancer" | hbase shell
sleep 10
cd /home/eclark/RC5/ycsb/
bin/ycsb load hbase -P workloads/workloada -p columnfamily=d -s -threads 32 -p 
recordcount=$RECORDS 2> /tmp/ycsb-workload-a-load-output.log
{code}


I've also seen the same stack and exception when trying to run ITBLL on this 
cluster.

> Client write path perf issues
> -----------------------------
>
>                 Key: HBASE-9775
>                 URL: https://issues.apache.org/jira/browse/HBASE-9775
>             Project: HBase
>          Issue Type: Bug
>          Components: Client
>    Affects Versions: 0.96.0
>            Reporter: Elliott Clark
>            Priority: Critical
>         Attachments: Charts Search   Cloudera Manager.png, short_ycsb.png
>
>
> Testing on larger clusters has not had the desired throughput increases.



--
This message was sent by Atlassian JIRA
(v6.1#6144)