[
https://issues.apache.org/jira/browse/HADOOP-2488?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12554455
]
stack commented on HADOOP-2488:
-------------------------------
Hmm... seems like checksum sizes are default of 512 in 0.15 branch so that
ain't it.
I did a clean checkout of trunk and spent more time on timings and my original
claim that TRUNK is 3times slower is incorrect: its *just* ~40% slower.
0.15 branch
{code}
[EMAIL PROTECTED] hbase]$ for i in 1 2 3 ; do ./bin/hbase
org.apache.hadoop.hbase.PerformanceEvaluation mapfile 1; done
07/12/26 19:08:11 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 19:08:11 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 19:08:36 INFO hbase.PerformanceEvaluation: Writing 100000 records took
24159ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 19:08:36 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 19:08:52 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 19:09:05 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 19:09:17 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 19:09:30 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 19:09:42 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 19:09:55 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 19:10:08 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 19:10:20 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 19:10:33 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 19:10:45 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 129836ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 19:10:45 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 19:10:47 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1717ms
07/12/26 19:10:48 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 19:10:48 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 19:11:12 INFO hbase.PerformanceEvaluation: Writing 100000 records took
23859ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 19:11:12 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 19:11:25 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 19:11:39 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 19:11:51 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 19:12:03 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 19:12:16 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 19:12:29 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 19:12:41 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 19:12:54 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 19:13:06 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 19:13:19 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 126779ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 19:13:19 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 19:13:20 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1769ms
07/12/26 19:13:21 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 19:13:21 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 19:13:45 INFO hbase.PerformanceEvaluation: Writing 100000 records took
23679ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 19:13:45 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 19:13:58 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 19:14:11 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 19:14:23 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 19:14:36 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 19:14:48 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 19:15:01 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 19:15:13 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 19:15:26 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 19:15:38 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 19:15:51 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 125860ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 19:15:51 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 19:15:52 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1836ms
{code}
TRUNK
{code}
[EMAIL PROTECTED] hbase]$ for i in 1 2 3 ; do ./bin/hbase
org.apache.hadoop.hbase.PerformanceEvaluation mapfile 1; done
07/12/26 18:50:17 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 18:50:17 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 18:50:41 INFO hbase.PerformanceEvaluation: Writing 100000 records took
24191ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 18:50:41 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 18:51:02 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 18:51:20 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 18:51:38 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 18:51:55 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 18:52:13 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 18:52:29 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 18:52:47 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 18:53:04 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 18:53:21 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 18:53:39 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 177692ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 18:53:39 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 18:53:41 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1749ms
07/12/26 18:53:42 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 18:53:42 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 18:54:06 INFO hbase.PerformanceEvaluation: Writing 100000 records took
23873ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 18:54:06 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 18:54:25 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 18:54:44 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 18:55:01 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 18:55:17 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 18:55:33 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 18:55:51 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 18:56:08 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 18:56:25 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 18:56:41 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 18:56:57 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 171167ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 18:56:57 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 18:56:59 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1789ms
07/12/26 18:57:00 WARN util.NativeCodeLoader: Unable to load native-hadoop
library for your platform... using builtin-java classes where applicable
07/12/26 18:57:00 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
hdfs://X.X.X:9123/user/?/performanceevaluation.mapfile
07/12/26 18:57:25 INFO hbase.PerformanceEvaluation: Writing 100000 records took
24076ms (Note: generation of keys and values is done inline and has been seen
to consume significant time: e.g. ~30% of cpu time
07/12/26 18:57:25 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
07/12/26 18:57:43 INFO hbase.PerformanceEvaluation: Read 10000
07/12/26 18:58:01 INFO hbase.PerformanceEvaluation: Read 20000
07/12/26 18:58:18 INFO hbase.PerformanceEvaluation: Read 30000
07/12/26 18:58:35 INFO hbase.PerformanceEvaluation: Read 40000
07/12/26 18:58:52 INFO hbase.PerformanceEvaluation: Read 50000
07/12/26 18:59:09 INFO hbase.PerformanceEvaluation: Read 60000
07/12/26 18:59:26 INFO hbase.PerformanceEvaluation: Read 70000
07/12/26 18:59:42 INFO hbase.PerformanceEvaluation: Read 80000
07/12/26 19:00:00 INFO hbase.PerformanceEvaluation: Read 90000
07/12/26 19:00:16 INFO hbase.PerformanceEvaluation: Reading 100000 random
records took 171260ms (Note: generation of random key is done in line and takes
a significant amount of cpu time: e.g 10-15%
07/12/26 19:00:16 INFO hbase.PerformanceEvaluation: Reading 100000 rows
sequentially
07/12/26 19:00:18 INFO hbase.PerformanceEvaluation: Reading 100000 records
serially took 1764ms
{code}
> Random reads in mapfile are 3times slower in TRUNK than they are in 0.15.x
> --------------------------------------------------------------------------
>
> Key: HADOOP-2488
> URL: https://issues.apache.org/jira/browse/HADOOP-2488
> Project: Hadoop
> Issue Type: Bug
> Reporter: stack
>
> Opening a mapfile on hdfs and then doing 100k random reads inside the open
> file takes 3 times longer to complete in TRUNK than same test run on 0.15.x.
> Random read performance is important to hbase.
> Serial reads and writes perform about the same in TRUNK and 0.15.x.
> Below are 3 runs done against 0.15 of the mapfile test that can be found in
> hbase followed by 2 runs done against TRUNK:
> 0.15 branch
> {code}
> [EMAIL PROTECTED] hbase]$ for i in 1 2 3; do ./bin/hbase
> org.apache.hadoop.hbase.PerformanceEvaluation mapfile 1; done
> 07/12/24 18:34:50 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 18:34:50 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXX:9123/user/?/performanceevaluation.mapfile
> 07/12/24 18:35:13 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 23644ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 18:35:13 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ....
> 07/12/24 18:37:23 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 129879ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 18:37:23 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 18:37:25 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1832ms
> 07/12/24 18:37:26 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 18:37:26 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXXX:9123/user/?/performanceevaluation.mapfile
> 07/12/24 18:37:50 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 24188ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 18:37:50 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ...
> 07/12/24 18:39:58 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 127879ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 18:39:58 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 18:40:00 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1787ms
> 07/12/24 18:40:01 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 18:40:01 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXX:9123/user/?/performanceevaluation.mapfile
> 07/12/24 18:40:24 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 23832ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 18:40:24 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ..
> 07/12/24 18:42:31 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 126954ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 18:42:31 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 18:42:33 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1766ms
> 07/12/24 17:24:25 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 17:24:25 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXXX:9123/user/?/performanceevaluation.mapfile
> 07/12/24 17:24:49 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 24181ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 17:24:49 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ...
> 07/12/24 17:26:59 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 129564ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 17:26:59 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 17:27:00 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1793ms
> {code}
> TRUNK
> {code}
> [EMAIL PROTECTED] hbase]$ ./bin/hbase
> org.apache.hadoop.hbase.PerformanceEvaluation mapfile 1
> 07/12/24 18:02:26 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 18:02:26 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXX3/user/?/performanceevaluation.mapfile
> 07/12/24 18:02:51 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 25500ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 18:02:51 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ....
> 07/12/24 18:11:11 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 500306ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 18:11:11 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 18:11:13 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1940ms
> [EMAIL PROTECTED] hbase]$ ./bin/hbase
> org.apache.hadoop.hbase.PerformanceEvaluation mapfile 1
> 07/12/24 18:12:46 WARN util.NativeCodeLoader: Unable to load native-hadoop
> library for your platform... using builtin-java classes where applicable
> 07/12/24 18:12:46 INFO hbase.PerformanceEvaluation: Writing 100000 rows to
> hdfs://XXXX/user/?/performanceevaluation.mapfile
> 07/12/24 18:13:12 INFO hbase.PerformanceEvaluation: Writing 100000 records
> took 25593ms (Note: generation of keys and values is done inline and has been
> seen to consume significant time: e.g. ~30% of cpu time
> 07/12/24 18:13:12 INFO hbase.PerformanceEvaluation: Reading 100000 random rows
> ...
> 07/12/24 18:22:16 INFO hbase.PerformanceEvaluation: Reading 100000 random
> records took 543992ms (Note: generation of random key is done in line and
> takes a significant amount of cpu time: e.g 10-15%
> 07/12/24 18:22:16 INFO hbase.PerformanceEvaluation: Reading 100000 rows
> sequentially
> 07/12/24 18:22:18 INFO hbase.PerformanceEvaluation: Reading 100000 records
> serially took 1786ms
> {code}
> Above was done on a small hdfs cluster of 4 machines with each a dfs node.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
