[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14997520#comment-14997520 ] Zhe Zhang commented on HDFS-8425: - bq. Per local test, it's 2.5x slower than repl. We need a faster codec HADOOP-11887 has just shipped. It's much faster than the current coder. I think it's very important to isolate different factors impacting performance. For this purpose, [~lirui] is working on HDFS-9345. With that we should be able to figure out any potential performance issue with the output logic (non-codec). Rui is also working on HDFS-8968, which is a more comprehensive benchmark for EC I/O. > [umbrella] Performance tuning, investigation and optimization for erasure > coding > > > Key: HDFS-8425 > URL: https://issues.apache.org/jira/browse/HDFS-8425 > Project: Hadoop HDFS > Issue Type: Sub-task >Affects Versions: HDFS-7285 >Reporter: GAO Rui > Attachments: testClientWriteReadFile_v1.pdf, > testdfsio-read-mbsec.png, testdfsio-write-mbsec.png > > > This {{umbrella}} jira aims to track performance tuning, investigation and > optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14984928#comment-14984928 ] Walter Su commented on HDFS-8425: - Thanks [~tfukudom]! The results looks good. I agree we should test read with some DN killed. But I'm afraid it won't be much different in the TestDFSIO. I've only tested writing. when I ran TestDFSIO, I found the throughput of ec is slightly better than repl. It's the same as [~tfukudom]'s tests. I open disk monitor and network monitor. The disk monitor shows that disk utilization often hits 100%. I think it's because we can use all the cpus of NodeManagers, so the bottleneck is disk/network io. It's useful because we can write ec files in batch. For example, converting multiple repl files to ec files. The speed of single client writing is constrained by coding speed. Per local test, it's 2.5x slower than repl. We need a faster codec. I think it's also important, right? But I'm not sure there's use-case is bounded by the speed of single client writing. Usually we write files using repl, and convert them to ec files later. How do you think? > [umbrella] Performance tuning, investigation and optimization for erasure > coding > > > Key: HDFS-8425 > URL: https://issues.apache.org/jira/browse/HDFS-8425 > Project: Hadoop HDFS > Issue Type: Sub-task >Affects Versions: HDFS-7285 >Reporter: GAO Rui > Attachments: testClientWriteReadFile_v1.pdf, > testdfsio-read-mbsec.png, testdfsio-write-mbsec.png > > > This {{umbrella}} jira aims to track performance tuning, investigation and > optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14984631#comment-14984631 ] Takuya Fukudome commented on HDFS-8425: --- Thanks for the comment, [~zhz]! The X-axis means the number of files(TestDFSIO's nrFiles parameter). And bq. And I guess you didn't kill and DN in read tests? Yes, you are right. I will do the read tests with failure situation later. Thank you! > [umbrella] Performance tuning, investigation and optimization for erasure > coding > > > Key: HDFS-8425 > URL: https://issues.apache.org/jira/browse/HDFS-8425 > Project: Hadoop HDFS > Issue Type: Sub-task >Affects Versions: HDFS-7285 >Reporter: GAO Rui > Attachments: testClientWriteReadFile_v1.pdf, > testdfsio-read-mbsec.png, testdfsio-write-mbsec.png > > > This {{umbrella}} jira aims to track performance tuning, investigation and > optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14981572#comment-14981572 ] Zhe Zhang commented on HDFS-8425: - Thanks [~tfukudom]! What does the X-axis show (10, 20, ..., 200)? And I guess you didn't kill and DN in read tests? > [umbrella] Performance tuning, investigation and optimization for erasure > coding > > > Key: HDFS-8425 > URL: https://issues.apache.org/jira/browse/HDFS-8425 > Project: Hadoop HDFS > Issue Type: Sub-task >Affects Versions: HDFS-7285 >Reporter: GAO Rui > Attachments: testClientWriteReadFile_v1.pdf, > testdfsio-read-mbsec.png, testdfsio-write-mbsec.png > > > This {{umbrella}} jira aims to track performance tuning, investigation and > optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14641195#comment-14641195 ] Zhe Zhang commented on HDFS-8425: - Moving system test JIRAs as follow-ons. Let me know if you have other opinions. Thanks! [umbrella] Performance tuning, investigation and optimization for erasure coding Key: HDFS-8425 URL: https://issues.apache.org/jira/browse/HDFS-8425 Project: Hadoop HDFS Issue Type: Sub-task Affects Versions: HDFS-7285 Reporter: GAO Rui Attachments: testClientWriteReadFile_v1.pdf This {{umbrella}} jira aims to track performance tuning, investigation and optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HDFS-8425) [umbrella] Performance tuning, investigation and optimization for erasure coding
[ https://issues.apache.org/jira/browse/HDFS-8425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14572086#comment-14572086 ] Li Bo commented on HDFS-8425: - I think encoding is one factor. More work can be done to enhance the efficiency. When the data buffers are full, we will suspend data writing and switch to generate parity buffers. I think they can work concurrently. I have created HDFS-8528 which tries to enhance the efficiency in this direction. [umbrella] Performance tuning, investigation and optimization for erasure coding Key: HDFS-8425 URL: https://issues.apache.org/jira/browse/HDFS-8425 Project: Hadoop HDFS Issue Type: Sub-task Affects Versions: HDFS-7285 Reporter: GAO Rui Attachments: testClientWriteReadFile_v1.pdf This {{umbrella}} jira aims to track performance tuning, investigation and optimization for erasure coding. -- This message was sent by Atlassian JIRA (v6.3.4#6332)