…performance that still stands in the way of InnoDB scalability is the trx_list scan on consistent…trx_list scan in MySQL 5.5 basically kills throughput on high concurrency. The picture changes … a notable drop in throughput starting from 256 threads. But Percona Server’s scalability is the same…
Post: trx descriptors: MySQL performance improvements in Percona Server 5.5.30-30.2
April 12, 2013 By 3 Comments
Post: Choosing innodb_buffer_pool_size
November 3, 2007 By 39 Comments
…, key_buffer, mysql threads, temporary tables, per thread sort buffer which can be allocated. There are also things like innodb additional memory pool…. You do not want OS to cache what Innodb is caching already. Innodb cache is more efficient compared to OS cache because… have memory usage spike you may have MySQL Server being killed by OS instead of temporary swapping few things out. Two…
Post: Ultimate MySQL variable and status reference list
January 8, 2011 By 4 Comments
….commanual innodb_status_fileblogpercona.commanual innodb_strict_modeblogpercona.commanual innodb_support_xablogpercona.commanual innodb_sync_spin_loopsblogpercona.commanual innodb_table_locksblogpercona.commanual innodb_thread_concurrencyblogpercona.commanual innodb_thread_sleep_delayblogpercona.commanual Innodb_truncated…
Post: Heikki Tuuri Innodb answers - Part I
October 26, 2007 By 19 Comments
…: Another possible area of optimization. I frequently see batch jobs killing server performance overtaking buffer pool. Though full table scan is… are not seeing ‘thread thrashing’ (lots of threads waiting for semaphores in SHOW INNODB STATUS), you can try to disable thread throttling completely by…
Post: The perils of InnoDB with Debian and startup scripts
January 28, 2009 By 13 Comments
… the other threads: —TRANSACTION 0 228527423, ACTIVE 0 sec, process no 30034, OS thread id 1159956816 starting index read, thread declared inside InnoDB 500 mysql tables in use 7, locked 0 MySQL thread id 6424… the top. The solution I chose in this case? mysql> KILL 7; Immediately afterward everything cleared up.
Post: MySQL 5.5.8 - in search of stability
January 3, 2011 By 11 Comments
…innodb_read_io_threads = 16 innodb_write_io_threads = 16 innodb_io_capacity=500 innodb_max_dirty_pages_pct = 60 innodb_purge_threads=1 innodb_adaptive_flushing=0 innodb… Basically, by lowering innodb_max_dirty_pages_pct, you are killing your throughput. When you disable innodb_doublewrite, you can …
Post: Percona Server 5.1.59-13.0
October 16, 2011 By Leave a Comment
… InnoDB Fake Changes When restarting a slave server in a replication environment, the process can be speed up by having prefetch threads… fast as in most cases there is nothing to do. InnoDB Kill Idle Transactions NOTE: Percona classes this feature as Beta and… killed. This prevents users from blocking purge by mistake. Block startup until LRU dump is loaded Added a new boolean option, –innodb…
Post: Percona Server 5.5.16-22.0
October 16, 2011 By 2 Comments
… InnoDB Fake Changes When restarting a slave server in a replication environment, the process can be speed up by having prefetch threads… fast as in most cases there is nothing to do. InnoDB Kill Idle Transactions NOTE: Percona classes this feature as Beta and… killed. This prevents users from blocking purge by mistake. Block startup until LRU dump is loaded Added a new boolean option, –innodb…
Post: When Does InnoDB Update Table Statistics? (And When It Can Bite)
October 6, 2011 By 7 Comments
… show up on slow query log until InnoDB again updates the statistics. But when does InnoDB perform the updates aside from the…? The 2 instances below are documented from the MySQL and InnoDB plugin’s manual: Metadata commands like SHOW INDEX, SHOW TABLE…: user1[user1] @ [172.20.6.1] # Thread_id: 10140441 Schema: db1 Last_errno: 0 Killed: 0 # Query_time: 12.277786 Lock_time…
Post: MySQL Partitioning - can save you or kill you
December 11, 2010 By 45 Comments
…) unsigned DEFAULT NULL, PRIMARY KEY (`id`), KEY `uu` (`uu`), ) ENGINE=InnoDB The access pattern to this table is to lookup data… in buffer pool completely so replication became CPU bound (single thread) instead of IO bound. You could celebrate but hey…. you… by UUID from 400 to 20 per second (from single thread). Decreasing number of partitions made replication less efficient but the…
