提供综合数据库运维服务与优化方案(不限Oracle MySQL PG GaussDB GoldenDB OceanBase等), 微信/Tel:(+86)134-365-60330
数据库oracle
Some scripts for showing the growth of tablespaces over time periods. to show the starting tablespace size and the end tablespace size, computing the tablespace growth for the time period.
内存管理是一个非常复杂的结构，在操作系统中用Virtual memory System进行内存管理，pagetable是中记录了内存的虚拟地址和物理地址的映射关系，当一个进程真正的访问数据是首先访问的是page table然后转向真实的地址，cpu中有固定的大小来缓存部份的pagetables,这块区域就是translation lookaside buffer(TLB).
今天查看AWR时，发现有个sql占用的cpu很高而且也并不复杂，top 5 event出现了cpu,read by other sessions,而且sql中再次看到的hint index,随后从生产环境中拿到绑定变量参数值进行本地调试，发现hint index 选择了错误的索引一致读为93万，去掉hint立即降低到了36，随后服务器压力将了下来
The log file switch – checkpoint incomplete event tells you that Oracle wants to reuse a redo log file, but the current checkpoint position is still in that log file. This results in redo generation being stopped until the log switch is done.
经常见到select * from table(function(args))的查询方法,对pl/sql 强大功能赞不绝口,比如PIPELINED 管道函数
例子
sometimes we use ALTER SESSION statement to set or modify any of the conditions or parameters that affect your connection to the database. The statement stays in effect until you disconnect from the database. we can query current system Initialization parameter from v$parameter view,but where are able to confirm that the session-level parameter modification happened?
ASM is the ability to add and remove disks/luns from a diskgroup with no down time and an automatic rebalancing of the data on the luns. This creates all kinds of possibilities.
11gr1起提供了Invisible Indexes,作用就是通过alter index xx invisible，对CBO隐身，我突然想到之前的一个情况，就是开发的在应用中滥用hint index，当时为了让hint index 无效，当时是选择index rename，现在看来如果是11g可以用invisible
awr 中有很多wr开头的表收集的数据非常有用，通过dba_hist开头的视图可以很直观的查看，也可以手动查询一些详细或定制自己所要的信息
when a long, running transaction has been rolled back, How Much Longer completion time of the rollback process? The questions are justified, because the transaction holds the locks and normal processing often suffers as the rollback progresses.
We use to import a dump file into a database today. impdp process was hang When the archive log destination disk space becomes full,I found this and drop older archivelog files from OS immediately. but disk space would be full as soon as I drop some files.and db was hang,I decided to terminate the impdp … Read more
有时需要清除schema里的所有对象，但又没有dba权限,不可以drop user xxx cascade; re-create user;可以写个plsql，循环删除所有对象，注意有些对象存在依赖关系，如外键、index 和table..,这里建个procedure，方便在开发库一堆有schema中运行。
如果primary db 丢失或不再需要的physical standby ，想把这些库(standby)利用起来，logical standby 本来就可以write open，11g ADG也可以APPLY LOG时open,但是read only，下面演示如果把off-line standby 转为primary,open read/write mode.
自定义生成一些报表，曲线数据从AWR 数据表。
CBO优化器的内部有两个重要的标准：selectivity and cardinality，cardinality 就CBO在处理完后返回的行数，selectivity是返回数据范围百分比，cardinality = (number of input rows)* selectivity,所以selectivity 的直接影响了CBO表访问路径
我是习惯在sqlplus中做一些操作，在linux中安装rlwrap 支持方向键，但如果多个用户在同一台机器上登录，每个人常用的命令不同，方向键查找起来很麻烦，so.那就把我们经常用的操作分类或以用户分类，设立独立的历史操作记录log
Hint提示是优化SQL的一种手段，但不应该放在首位，记的国外有位大师说过顺序应该是学写SQL，写更好SQL，学写ORACLE SQL，写更好的ORACLE SQL，在了解原理且CBO没有走理想路线的情况下再去指引CBO。
有时需要参考一些隐藏参数值了解ORACLE内部的限制,在sqlplus 中使用show parameter 显示隐藏参数..
Session cursor caching does not mean that server process caches the whole SQL area in the PGA.It just means that Oracle does not release the cursor object when the same statement was executed more than 2 times in the same process…
今天看到一个案例，从10G升级到11GR2后cpu负载迅速上升，很快达到100%,通过OWI显示出现大量“Cursor: mutex S”事件，mutex S是发生在cursor cache上的序列化mutex