Innodb: 自动开启打印show engine status到err日志

2023年 4月 29日 35.2k 0

一、问题描述 为什么我的err日志里面有大量的show engine innodb status的记录,我自己并没有开启innodb_status_output参数。 二、问题分析 通过查看日志,发现如下输出: 2019-03-21T17:00:02.375231

一、问题描述

为什么我的err日志里面有大量的show engine innodb status的记录,我自己并没有开启innodb_status_output参数。

二、问题分析

通过查看日志,发现如下输出: 2019-03-21T17:00:02.375231Z 1230497 [Warning] InnoDB: Difficult to find free blocks in the buffer pool (338 search iterations)! 0 failed attempts to flush a page! Consider increasing the buffer pool size. It is also possible that in your Unix version fsync is very slow, or completely frozen inside the OS kernel. Then upgrading to a newer version of your operating system may help. Look at the number of fsyncs in diagnostic info below. Pending flushes (fsync) log: 0; buffer pool: 0. 1446962050 OS file reads, 545881917 OS file writes, 376257282 OS fsyncs. Starting InnoDB Monitor to print further diagnostics to the standard output.

日志也写得很清楚。应该是free block不够了Innodb自动开启了。但是我们需要源码验证一下。

三、源码验证

在源码的buf_LRU_handle_lack_of_free_blocks函数中我们看到如下: if ((current_ms > started_ms + 2000)        && (current_ms > last_printout_ms + 2000)        && srv_buf_pool_old_size == srv_buf_pool_size) {        ib::warn() << "Difficult to find free blocks in the buffer pool"            " (" << n_iterations << " search iterations)! "              << flush_failures << " failed attempts to"            " flush a page! Consider increasing the buffer pool"            " size. It is also possible that in your Unix version"            " fsync is very slow, or completely frozen inside"            " the OS kernel. Then upgrading to a newer version"            " of your operating system may help. Look at the"            " number of fsyncs in diagnostic info below."            " Pending flushes (fsync) log: "              << fil_n_pending_log_flushes              << "; buffer pool: "              << fil_n_pending_tablespace_flushes              << ". " << os_n_file_reads << " OS file reads, "              << os_n_file_writes << " OS file writes, "              << os_n_fsyncs              << " OS fsyncs. Starting InnoDB Monitor to print"            " further diagnostics to the standard output.";        last_printout_ms = current_ms;        *mon_value_was = srv_print_innodb_monitor;        *started_monitor = true;        srv_print_innodb_monitor = true;        os_event_set(srv_monitor_event);

这里不仅打印出了日志同时设置了参数srv_print_innodb_monitor = true; 并且开始os_event_set(srv_monitor_event);开启了monitor打印线程。那我们看看srv_print_innodb_monitor 对应什么参数呢。如下: static MYSQL_SYSVAR_BOOL(status_output, srv_print_innodb_monitor,  PLUGIN_VAR_OPCMDARG, "Enable InnoDB monitor output to the error log.",  NULL, innodb_status_output_update, FALSE);

实际上就是innodb_status_output被自动开了。当然如果查看调用可以在上层函数buf_LRU_get_free_block中查看到调用,实际上就是在free list找不到空闲的block的时候会做输出。buf_LRU_get_free_block还包含了一个块的分配流程大约如下,可自行参考:

 If there is a block in the free list, take it .如果这里找不到就会自动开启innodb_status_output
 If no block was in the free list, search from the end of the LRU list and try to free a block there.
 No free block was found: try to flush the LRU list.

相关文章

最新发布!MySQL 9.0 的向量 (VECTOR) 类型文档更新
国产数据库中级认证HCIP-openGauss经验分享
保障数据完整性与稳定性:数据库一致性
OceanBase 里的 DDL 超时时间
OceanBase v3.1.x 将不再更新版本 | 社区月报2024.6
openGauss Developer Day 2024 | SIG组工作会议亮点回看!

发布评论