orzdba 监控获取/输出全解

orzdba好用的不要不要的

threads
  • run:Threads_running
  • con:Threads_connected
  • cre:Threads_created
  • cac:Threads_cached
------threads------
 run  con  cre  cac|
   0    0    0    0|
 200  202    0    0|
 201  202    0    0|
 201  202    0    0|
 201  202    0    0|
innodb_hit
  • lor : Innodb_buffer_pool_read_requests [innodb进行逻辑读的数量。单位是次数]
  • hit : (Innodb_buffer_pool_read_requests - Innodb_buffer_pool_reads[读磁盘次数]) / Innodb_buffer_pool_read_requests * 100%
        -Hit%-
|     lor    hit|
|       0 100.00|
|  385098  98.00|
|  392831  98.01|
innodb_rows
  • innodb_rows_inserted_diff : innodb_rows_inserted
  • innodb_rows_updated_diff :innodb_rows_updated
  • innodb_rows_deleted_diff :innodb_rows_deleted
  • innodb_rows_read_diff :innodb_rows_read
---innodb rows status---
  ins   upd   del   read|
    0     0     0      0|
 1717  3430  1718 715111|
 1728  3439  1723 717916|
innodb_data
  • reads:Innodb_data_reads 次数
  • writes:innodb_data_writes 次数
  • read:Innodb_data_read 字节
  • written:Innodb_data_written 字节
-----innodb data status----
 reads writes  read written|
     0      0      0      0|
  8220   5043 128.4m 106.3m|
  8091   5139 126.4m 108.6m|
  8089   5148 126.4m 109.3m|
innodb_log
  • fsyncs:Innodb_os_log_fsyncs 记录往磁盘上写日志的次数
  • written:innodb_os_log_written 日志刷盘的字节数,如果在commit不怎么变化的情况下,这个值出现暴增,说明系统出现大事务了(处理:kill线程,必要情况kill掉mysql进程)。写入redo logfile中的字节量
--innodb log--
fsyncs written|
     0       0|
  1701    3.9m|
  1708    4.1m|
  1716    3.9m|
innodb_status

获取方式:show engine innodb status

  • his

    • list : 取值方式:history list length innodb的history list长度代表了有多少undo日志还没有被清理掉。如果发现history list的长度越大,要么就是实例的复杂非常高,要么就是可能有大查询,或者事务没提交,导致Undo log无法分析
  • log(byte)

    • uflush:= (Log sequence number)-(Log flushed up to) :没有刷新到redolog的文件大小
    • uckpt:= (Log sequence number)-(Last checkpoint at) :没有落盘的redo大小
  • read

    • view :read views open inside InnoDB。有多少read view被打开了,一个read view是一致性保证的MVCC “snapshot”
  • query

    • inside :有多少线程在Innodb内核
    • que :队列长度
  his --log(byte)--  read ---query---
 list uflush  uckpt  view inside  que|
  566     1k  891.9m  191     4   193|
  757     2k  894.9m  189     5   194|
  529     3k  898.0m  193     3   194|
  427    503  880.6m  190     1   193|


---
LOG
(Innodb 事务日志相关信息,包括当前的日志序列号(Log sequence number),已经刷新同步到那个序列号,最近的check point到那个序列号了。除此之外,还显示了系统从启动到现在已经做了多少次check point,多少次日志刷新。)
---
(注:小括号为官方解释。)
Log sequence number 2560255(当前的日志序列号)
Log flushed up to   2560255(刷新到日志重做日志文件的lsn)
Pages flushed up to 2560255(写入磁盘的脏页的lsn。记录在checkpoint中)
Last checkpoint at 2560246(刷新到磁盘的lsn)

 --------------
 ROW OPERATIONS
 --------------
 2 queries inside InnoDB, 0 queries in queue
 2 read views open inside InnoDB
 Main thread process no. 7969, id 1191348544, state: sleeping
 Number of rows inserted 287921794, updated 733493588, deleted 30775703, read 2351464150250
 5.10 inserts/s, 29.38 updates/s, 0.02 deletes/s, 51322.87 reads/s

你可能感兴趣的:(orzdba 监控获取/输出全解)