Comments on: Group commit and real fsync http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/ Everything about MySQL Performance Sat, 21 Nov 2009 05:23:57 -0800 http://wordpress.org/?v=2.8.4 hourly 1 By: Pretending to fix broken group commit | MySQL Performance Blog http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-462492 Pretending to fix broken group commit | MySQL Performance Blog Tue, 03 Feb 2009 03:50:11 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-462492 [...] write performance in concurrent load drops down significantly. We wrote also about it before, see http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/ and http://www.mysqlperformanceblog.com/2006/05/19/group-commit-and-xa/ The problem is the InnoDB [...] [...] write performance in concurrent load drops down significantly. We wrote also about it before, see http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/ and http://www.mysqlperformanceblog.com/2006/05/19/group-commit-and-xa/ The problem is the InnoDB [...]

]]> By: Okulov Vitaliy http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-322040 Okulov Vitaliy Thu, 03 Jul 2008 08:02:03 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-322040 My benchmark with sysbench: sysbench v0.4.8: multi-threaded system evaluation benchmark Running the test with following options: Number of threads: 1 Extra file open flags: 0 1 files, 16Kb each 16Kb total file size Block size 16Kb Number of random requests for random IO: 10000 Read/Write ratio for combined random IO test: 1.50 Periodic FSYNC enabled, calling fsync() each 1 requests. Calling fsync() at the end of test, Enabled. Using synchronous I/O mode Doing random write test Threads started! Done. Operations performed: 0 Read, 10000 Write, 10000 Other = 20000 Total Read 0b Written 156.25Mb Total transferred 156.25Mb (83.458Mb/sec) 5341.28 Requests/sec executed Test execution summary: total time: 1.8722s total number of events: 10000 total time taken by event execution: 0.1558 per-request statistics: min: 0.0000s avg: 0.0000s max: 0.0000s approx. 95 percentile: 0.0000s Threads fairness: events (avg/stddev): 10000.0000/0.00 execution time (avg/stddev): 0.1558/0.00 Hardware: 2x5130 2.3 Ghz, 6 Gb, 237 RAID 10 Adaptec 2130SLP BBU. My benchmark with sysbench:

sysbench v0.4.8: multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
1 files, 16Kb each
16Kb total file size
Block size 16Kb
Number of random requests for random IO: 10000
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 1 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random write test
Threads started!
Done.

Operations performed: 0 Read, 10000 Write, 10000 Other = 20000 Total
Read 0b Written 156.25Mb Total transferred 156.25Mb (83.458Mb/sec)
5341.28 Requests/sec executed

Test execution summary:
total time: 1.8722s
total number of events: 10000
total time taken by event execution: 0.1558
per-request statistics:
min: 0.0000s
avg: 0.0000s
max: 0.0000s
approx. 95 percentile: 0.0000s

Threads fairness:
events (avg/stddev): 10000.0000/0.00
execution time (avg/stddev): 0.1558/0.00

Hardware: 2×5130 2.3 Ghz, 6 Gb, 237 RAID 10 Adaptec 2130SLP BBU.

]]> By: MySQL Blog: Alexander Rubin’s Blog http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-200200 MySQL Blog: Alexander Rubin’s Blog Tue, 20 Nov 2007 21:41:10 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-200200 [...] originally performed InnoDB scalability tests for MySQL 5.0 to show broken group commit [...] [...] originally performed InnoDB scalability tests for MySQL 5.0 to show broken group commit [...]

]]> By: Nilnandan Joshi http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-25053 Nilnandan Joshi Thu, 21 Dec 2006 06:57:25 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-25053 Hi, I want to optimise the ORDER BY in MySQL how can i???? pls help me. I run 1 query with GROUP BY without ORDER BY then it takes 0.0417 seconds but when i add ORDER BY then it takes 12 seconds..so, can i use any alternative of ORDER BY..???? Thanks Nilnandan Joshi Hi,

I want to optimise the ORDER BY in MySQL how can i???? pls help me.
I run 1 query with GROUP BY without ORDER BY then it takes 0.0417 seconds but when i add ORDER BY then it takes 12 seconds..so, can i use any alternative of ORDER BY..????

Thanks

Nilnandan Joshi

]]> By: 分组提交和实时fsync at 周俭 Blog http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-4677 分组提交和实时fsync at 周俭 Blog Mon, 02 Oct 2006 00:16:33 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-4677 [...] 原文作者： Peter Zaitsev原文来源： http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/译者：叶金荣（Email:），转载请注明译者和出处，并且不能用于商业用途，违者必究。 [...] [...] 原文作者： Peter Zaitsev原文来源： http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/译者：叶金荣（Email:），转载请注明译者和出处，并且不能用于商业用途，违者必究。 [...]

]]> By: MySQL Performance Blog » Watch out for Marketing benchmarks http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-1968 MySQL Performance Blog » Watch out for Marketing benchmarks Wed, 23 Aug 2006 17:50:16 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-1968 [...] Of course I would not say it is normal and nothing can be done in all cases. There are performance bugs, some of them as Innodb Scalability were known long before MySQL 5.0 was stable but were not fixed, others, such as broken group commit could have been handled better. [...] [...] Of course I would not say it is normal and nothing can be done in all cases. There are performance bugs, some of them as Innodb Scalability were known long before MySQL 5.0 was stable but were not fixed, others, such as broken group commit could have been handled better. [...]

]]> By: peter http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-45 peter Tue, 23 May 2006 17:11:12 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-45 Jay, There is no mystery in this problem. It is reported as bug as you see and Heikki needs to fix it. However the problem is it is not tiny little bug it is feature which was not implemented in 5.0. Simply the new code had to be written to make that work with XA but it never was done. So now it is kind of dangerous to fix it in 5.0 Jay,

There is no mystery in this problem. It is reported as bug as you see and Heikki needs to fix it. However the problem is it is not tiny little bug it is feature which was not implemented in 5.0. Simply the new code had to be written to make that work with XA but it never was done. So now it is kind of dangerous to fix it in 5.0

]]> By: Jay Pipes http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-44 Jay Pipes Tue, 23 May 2006 15:48:56 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-44 Hi Peter! Wonderful blog entries, as usual! Hope your move is going OK. Do you have any insider updates on the group commit bug? I checked the worklog and found very little going on. Do you know if Innobase has found the root cause of the issue? Cheers, and talk with you soon! Jay Hi Peter!

Wonderful blog entries, as usual! Hope your move is going OK. Do you have any insider updates on the group commit bug? I checked the worklog and found very little going on. Do you know if Innobase has found the root cause of the issue?

Cheers, and talk with you soon!

Jay

]]> By: MySQL Performance Blog » Blog Archive » Group commit and XA http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/comment-page-1/#comment-36 MySQL Performance Blog » Blog Archive » Group commit and XA Fri, 19 May 2006 19:59:21 +0000 http://www.mysqlperformanceblog.com/2006/05/03/group-commit-and-real-fsync/#comment-36 [...] Group commit and XA Returning to post Group commit and real fsync I made several experiments: I ran sysbench update_key benchmarks without –log-bin, with –log-bin, and with –log-bin and –innodb-support-xa = 0 (default value is 1). Results (in transactions / sec) threads 1     2            3 1  1218.68  614.94  1010.44 4  2686.36  667.77  1162.60 16  3993.59  666.14  1161.56 64  3630.55  665.18  1151.36 (column 1 - without log-bin  2 - with –log-bin  3 - with –log-bin and –innodb_support-xa=0 ) As you see with –innodb_support-xa=0 the result is almost by 2 times better than without (if –log-bin is used). So, if you are using –log-bin with innodb tables it makes sense to set –innodb_support-xa=0 [...] [...] Group commit and XA Returning to post Group commit and real fsync I made several experiments: I ran sysbench update_key benchmarks without –log-bin, with –log-bin, and with –log-bin and –innodb-support-xa = 0 (default value is 1). Results (in transactions / sec) threads 1     2            3 1  1218.68  614.94  1010.44 4  2686.36  667.77  1162.60 16  3993.59  666.14  1161.56 64  3630.55  665.18  1151.36 (column 1 – without log-bin  2 – with –log-bin  3 – with –log-bin and –innodb_support-xa=0 ) As you see with –innodb_support-xa=0 the result is almost by 2 times better than without (if –log-bin is used). So, if you are using –log-bin with innodb tables it makes sense to set –innodb_support-xa=0 [...]

]]>