[
https://issues.apache.org/jira/browse/TS-1405?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13631087#comment-13631087
]
Zhao Yongming commented on TS-1405:
-----------------------------------
after I have some stress testing in our testing lab, with jtest
the box is 24 core(logically), with dule-10GE nic, all ssd system
each TS set to 20ET_NET threads
and 4 client stressing one server, on each client, I start 6 jtest with:
{code}
screen jtest -P xxx.xxx.xxx.xxx -S ts.cn -s 9990 -z 0.99 -D xxx -c 30
{code}
that is a 99% hit condition
and the numbers show that in full load, the patch may provide better
performance in response time(12% less) but used a litter more cpu(2% more)
the current master:
{code}
Time -----------------------cpu----------------------
--------------------ts------------------
Time user sys wait hirq sirq util qps cons
Bps rt rpc
14/04/13-00:21 23.81 10.66 0.36 0.00 34.06 68.53 92.0K 18.4K
1.3G 3.38 5.00
14/04/13-00:22 23.72 10.61 0.36 0.00 34.18 68.50 92.0K 18.4K
1.3G 3.37 5.00
14/04/13-00:23 23.55 10.60 0.31 0.00 33.65 67.80 91.0K 18.2K
1.3G 3.44 5.00
14/04/13-00:24 23.72 10.68 0.35 0.00 34.13 68.53 92.1K 18.4K
1.3G 3.35 5.00
14/04/13-00:25 23.75 10.68 0.35 0.00 34.11 68.55 92.2K 18.4K
1.3G 3.33 5.00
14/04/13-00:26 23.51 10.53 0.33 0.00 33.66 67.71 90.9K 18.2K
1.3G 3.53 5.00
14/04/13-00:27 23.81 10.63 0.35 0.00 34.00 68.44 91.9K 18.4K
1.3G 3.46 5.00
14/04/13-00:28 23.79 10.64 0.36 0.00 34.04 68.46 92.2K 18.4K
1.3G 3.37 5.00
14/04/13-00:29 23.72 10.67 0.34 0.00 33.93 68.32 91.9K 18.4K
1.3G 3.34 5.00
14/04/13-00:30 23.95 10.63 0.34 0.00 34.18 68.76 92.5K 18.5K
1.3G 3.31 5.00
{code}
the current master with patch version v11jp:
{code}
Time -----------------------cpu----------------------
--------------------ts------------------
Time user sys wait hirq sirq util qps cons
Bps rt rpc
14/04/13-00:22 25.20 10.55 0.25 0.00 34.12 69.87 91.9K 18.4K
1.3G 2.97 5.00
14/04/13-00:23 25.36 10.59 0.26 0.00 34.30 70.25 92.4K 18.5K
1.3G 2.98 5.00
14/04/13-00:24 25.51 10.57 0.26 0.00 34.23 70.30 92.1K 18.4K
1.3G 2.99 5.00
14/04/13-00:25 25.12 10.54 0.26 0.00 34.01 69.66 91.5K 18.3K
1.3G 2.98 5.00
14/04/13-00:26 25.33 10.57 0.25 0.00 34.33 70.23 92.2K 18.4K
1.3G 2.93 5.00
14/04/13-00:27 25.40 10.64 0.26 0.00 34.16 70.20 92.4K 18.5K
1.3G 2.94 5.00
14/04/13-00:28 25.25 10.53 0.26 0.00 33.94 69.72 91.6K 18.3K
1.3G 3.01 5.00
14/04/13-00:29 25.34 10.63 0.25 0.00 34.16 70.14 92.4K 18.5K
1.3G 2.93 5.00
14/04/13-00:30 25.41 10.55 0.26 0.00 34.21 70.17 92.3K 18.5K
1.3G 2.96 5.00
14/04/13-00:31 25.33 10.56 0.26 0.00 34.11 69.99 91.8K 18.4K
1.3G 2.97 5.00
14/04/13-00:32 25.42 10.62 0.26 0.00 34.29 70.32 92.3K 18.5K
1.3G 2.92 5.00
{code}
> apply time-wheel scheduler about event system
> ----------------------------------------------
>
> Key: TS-1405
> URL: https://issues.apache.org/jira/browse/TS-1405
> Project: Traffic Server
> Issue Type: Improvement
> Components: Core
> Affects Versions: 3.2.0
> Reporter: Bin Chen
> Assignee: Bin Chen
> Fix For: 3.3.2
>
> Attachments: linux_time_wheel.patch, linux_time_wheel_v10jp.patch,
> linux_time_wheel_v11jp.patch, linux_time_wheel_v2.patch,
> linux_time_wheel_v3.patch, linux_time_wheel_v4.patch,
> linux_time_wheel_v5.patch, linux_time_wheel_v6.patch,
> linux_time_wheel_v7.patch, linux_time_wheel_v8.patch,
> linux_time_wheel_v9jp.patch
>
>
> when have more and more event in event system scheduler, it's worse. This is
> the reason why we use inactivecop to handler keepalive. the new scheduler is
> time-wheel. It's have better time complexity(O(1))
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
