Comments on: Tuning MySQL/InnoDB Flushing for a Write-Intensive Workload https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/ Sat, 20 Jan 2024 00:26:43 +0000 hourly 1 https://wordpress.org/?v=6.5.2 By: Yves Trudeau https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/#comment-10972426 Mon, 29 Jun 2020 18:15:12 +0000 https://www.percona.com/blog/?p=67999#comment-10972426 Hi Fred, 8.0.20 was released after this post, at first glance, the new doublewrite buffer seems to be rather similar to the Percona server parallel doublewrite buffer. I’ll check in more detail when I have a minute.

]]> By: lefred https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/#comment-10972378 Mon, 22 Jun 2020 09:41:36 +0000 https://www.percona.com/blog/?p=67999#comment-10972378 Hi Yves, Francisco,

You wrote that with 4 cleaner threads, InnoDB is able to flush at a very high rate but without parallel doublewrite buffers , the doublewrite buffer will bottleneck before the cleaner threads.

Have you tested with MySQL 8.0.20 (https://dev.mysql.com/doc/refman/8.0/en/innodb-parameters.html#sysvar_innodb_doublewrite_files).

Now if you think that the default of 2 should be 4, like the numbers of innodb_page_cleaners, please let us know 😉

Cheers,

]]> By: Mark Callaghan https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/#comment-10972250 Tue, 19 May 2020 15:20:11 +0000 https://www.percona.com/blog/?p=67999#comment-10972250 The problem with RocksDB is too many options that can be set. The problem isn’t too many options that should be set. It would be great if there were simple way to partition the option name space into “can set” vs “should set” to reduce confusion. An old post from me on this is http://smalldatum.blogspot.com/2018/09/5-things-to-set-when-configuring.html

From the above it isn’t clear to me whether most of these options are “can set” or “should set”. Perhaps that is content for the next blog post by you. Which options should I set?

]]> By: Yves Trudeau https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/#comment-10972240 Mon, 18 May 2020 13:08:26 +0000 https://www.percona.com/blog/?p=67999#comment-10972240 Exactly, innodb_io_capacity_max is more important than innodb_io_capacity. I recently worked with MyRocks, may I suggest your ease of tuning goes with your familiarity with then engine. It ended up working great but let’s say, I had an opportunity to experiment with column families… There are still a few things I don’t understand enough with MyRocks. I may blog about that in the near future.

]]> By: Mark Callaghan https://www.percona.com/blog/tuning-mysql-innodb-flushing-for-a-write-intensive-workload/#comment-10972233 Sat, 16 May 2020 16:30:25 +0000 https://www.percona.com/blog/?p=67999#comment-10972233 I was confused after reading the official docs for these parameters. I remain confused after reading this. What started my confusion was the statement that innodb_io_capacity doesn’t need to be changed. I assume this implies that innodb_io_capacity_max should still be set and I would be less confused had you added that.

Regardless, InnoDB has way too many tuning options in this space. It is starting to make MyRocks & RocksDB look easy to tune.

]]>