Oracle的数据库日志称为Redo log所有数据改变都记录Redo log可以用于修复受损的数据库Redo log 是用于恢复和一个高级特性的重要数据一个redo条目包含了相应操作导致的数据库变化的所有信息所有redo条目最终都要被写入redo文件中去
Redo log buffer是为了避免Redo文件IO导致性能瓶颈而在sga中分配出的一块内存一个redo条目首先在用户内存(PGA)中产生然后由oracle服务进程拷贝到log buffer中当满足一定条件时再由LGWR进程写入redo文件由于log buffer是一块共享内存为了避免沖突它是受到redo allocation latch保护的每个服务进程需要先获取到该latch才能分配redo buffer因此在高并发且数据修改频繁的oltp系统中我们通常可以观察到redo allocation latch的等待Redo写入redo buffer的整个过程如下
在PGA中生产Redo Enrey > 服务进程获取Redo Copy latch(存在多个CPU_COUNT*) > 服务进程获取redo allocation latch(仅个) > 分配log buffer > 释放redo allocation latch > 将Redo Entry写入Log Buffer > 释放Redo Copy latch;
shared strand
为了减少redo allocation latch等待在oracle 中引入了log buffer的并行机制其基本原理就是将log buffer划分为多个小的buffer这些小的buffer被成为strand(为了和之后出现的private strand区别它们被称之为shared strand)每一个strand受到一个单独redo allocation latch的保护多个shared strand的出现使原来序列化的redo buffer分配变成了并行的过程从而减少了redo allocation latch等待
shared strand的初始数据量是由参数log_parallelism控制的;在g中该参数成为隐含参数并新增参数_log_parallelism_max控制shared strand的最大数量;_log_parallelism_dynamic则控制是否允许shared strand数量在_log_parallelism和_log_parallelism_max之间动态变化
HELLODBACOM>select namksppinm valKSPPSTVL namksppdesc
from sysx$ksppi nam
sysx$ksppsv val
where namindx = valindx
AND namksppinm LIKE _%
AND upper(namksppinm) LIKE %LOG_PARALLE%;
KSPPINM KSPPSTVL KSPPDESC
_log_parallelism Number of log buffer strands
_log_parallelism_max Maximum number of log buffer strands
_log_parallelism_dynamic TRUE Enable dynamic strands
每一个shared strand的大小 = log_buffer/(shared strand数量)strand信息可以由表x$kcrfstrand查到(包含shared strand和后面介绍的private strandg以后存在)
HELLODBACOM>select indxstrand_size_kcrfa from x$kcrfstrand where last_buf_kcrfa != ;
INDX STRAND_SIZE_KCRFA
HELLODBACOM>show parameter log_buffer
NAME TYPE VALUE
log_buffer integer
关于shared strand的数量设置个cpu之内最大默认为当系统中存在redo allocation latch等待时每增加个cpu可以考虑增加个strand最大不应该超过并且_log_parallelism_max不允许大于cpu_count
注意在g中参数_log_parallelism被取消shared strand数量由_log_parallelism_max_log_parallelism_dynamic和cpu_count控制
Private strand
为了进一步降低redo buffer沖突在g中引入了新的strand机制Private strandPrivate strand不是从log buffer中划分的而是在shared pool中分配的一块内存空间
HELLODBACOM>select * from V$sgastat where name like %strand%;
POOL NAME BYTES
shared pool private strands
HELLODBACOM>select indxstrand_size_kcrfa from x$kcrfstrand where last_buf_kcrfa = ;
INDX STRAND_SIZE_KCRFA
Private strand的引入为Oracle的Redo/Undo机制带来很大的变化每一个Private strand受到一个单独的redo allocation latch保护每个Private strand作为私有的strand只会服务于一个活动事务获取到了Private strand的用户事务不是在PGA中而是在Private strand生成Redo当flush private strand或者commit时Private strand被批量写入log文件中如果新事务申请不到Private strand的redo allocation latch则会继续遵循旧的redo buffer机制申请写入shared strand中事务是否使用Private strand可以由x$ktcxb的字段ktcxbflg的新增的第位鑒定
HELLODBACOM>select decode(bitand(ktcxbflg )) used_private_strand count(*)
from x$ktcxb
where bitand(ksspaflg ) !=
and bitand(ktcxbflg ) !=
group by bitand(ktcxbflg );
USED_PRIVATE_STRAND COUNT(*)
对于使用Private strand的事务无需先申请Redo Copy Latch也无需申请Shared Strand的redo allocation latch而是flush或commit是批量写入磁盘因此减少了Redo Copy Latch和redo allocation latch申请/释放次数也减少了这些latch的等待从而降低了CPU的负荷过程如下
事务开始 > 申请Private strand的redo allocation latch (申请失败则申请Shared Strand的redo allocation latch) > 在Private strand中生产Redo Enrey > Flush/Commit > 申请Redo Copy Latch > 服务进程将Redo Entry批量写入Log File > 释放Redo Copy Latch > 释放Private strand的redo allocation latch
注意对于未能获取到Private strand的redo allocation latch的事务在事务结束前即使已经有其它事务释放了Private strand也不会再申请Private strand了
每个Private strand的大小为Kg中shared pool中的Private strands的大小就是活跃会话数乘以K而g中在shared pool中需要为每个Private strand额外分配k的管理空间即数量*k
g:
SQL> select * from V$sgastat where name like %strand%;
POOL NAME BYTES
shared pool private strands
HELLODBACOM>select trunc(value * KSPPSTVL / ) * *
from (select value from v$parameter where name = transactions) a
(select valKSPPSTVL
from sysx$ksppi nam sysx$ksppsv val
where namindx = valindx
AND namksppinm = _log_private_parallelism_mul) b;
TRUNC(VALUE*KSPPSTVL/)**
g:
HELLODBACOM>select * from V$sgastat where name like %strand%;
POOL NAME BYTES
shared pool private strands
HELLODBACOM>select trunc(value * KSPPSTVL / ) * ( + ) *
from (select value from v$parameter where name = transactions) a
(select valKSPPSTVL
from sysx$ksppi nam sysx$ksppsv val
where namindx = valindx
AND namksppinm = _log_private_parallelism_mul) b;
TRUNC(VALUE*KSPPSTVL/)*(+)*
Private strand的数量受到个方面的影响logfile的大小和活跃事务数量
参数_log_private_mul指定了使用多少logfile空间预分配给Private strand默认为我们可以根据当前logfile的大小(要除去预分配给log buffer的空间)计算出这一约束条件下能够预分配多少个Private strand
HELLODBACOM>select bytes from v$log where status = CURRENT;
BYTES
HELLODBACOM>select trunc(((select bytes from v$log where status = CURRENT) (select to_number(value) from v$parameter where name = log_buffer))*
(select to_number(valKSPPSTVL)
from sysx$ksppi nam sysx$ksppsv val
where namindx = valindx
AND namksppinm = _log_private_mul) / / )
as calculated private strands
from dual;
calculated private strands
HELLODBACOM>select count() actual private strands from x$kcrfstrand where last_buf_kcrfa = ;
actual private strands
当logfile切换后(和checkpoint一样切换之前必须要将所有Private strand的内容flush到logfile中因此我们在alert log中可能会发现日志切换信息之前会有这样的信息Private strand flush not complete这是可以被忽略的)会重新根据切换后的logfile的大小计算对Private strand的限制
HELLODBACOM>alter system switch logfile;
System altered
HELLODBACOM>select bytes from v$log where status = CURRENT;
BYTES
HELLODBACOM>select trunc(((select bytes from v$log where status = CURRENT) (select to_number(value) from v$parameter where name = log_buffer))*
(select to_number(valKSPPSTVL)
from sysx$ksppi nam sysx$ksppsv val
where namindx = valindx
AND namksppinm = _log_private_mul) / / )
as calculated private strands
from dual;
calculated private strands
HELLODBACOM>select count() actual private strands from x$kcrfstrand where last_buf_kcrfa = ;
actual private strands
参数_log_private_parallelism_mul用于推算活跃事务数量在最大事务数量中的百分比默认为Private strand的数量不能大于活跃事务的数量
HELLODBACOM>show parameter transactions
NAME TYPE VALUE
transactions integer
transactions_per_rollback_segment integer
HELLODBACOM>select trunc((select to_number(value) from v$parameter where name = transactions) *
(select to_number(valKSPPSTVL)
from sysx$ksppi nam sysx$ksppsv val
where namindx = valindx
AND namksppinm = _log_private_parallelism_mul) / )
as calculated private strands
from dual;
calculated private strands
HELLODBACOM>select count() actual private strands from x$kcrfstrand where last_buf_kcrfa = ;
actual private strands
注在预分配Private strand时会选择上述个条件限制下最小一个数量但相应的shared pool的内存分配和redo allocation latch的数量是按照活跃事务数预分配的
因此如果logfile足够大_log_private_parallelism_mul与实际活跃进程百分比基本相符的话Private strand的引入基本可以消除redo allocation latch的争用问题
