oracle的控制文件深入理解（一）
Oracle的control文件
这里我们要实验的是数据库在open的状态下，破快控制文件，数据库会不会down
查看controlfiles的路径
SQL> show parameter control_files
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
control_files string /u01/app/oracle/oradata/orcl/control01.ctl, /u01/app/oracle/fast_recovery_area/orcl/control02.ctl
破坏控制文件
[oracle@togogo ~]$ cat /dev/null > /u01/app/oracle/oradata/orcl/control01.ctl
[oracle@togogo ~]$ cat /dev/null > /u01/app/oracle/fast_recovery_area/orcl/control02.ctl
验证数据库的状态
SQL> select * from dual;
D
-
X
#### 做检查点，切换redo log
SQL> alter system checkpoint;
System altered.
SQL> alter system switch logfile;
System altered.
SQL> alter system switch logfile;
System altered.
#### check alert log
Sat Feb 10 20:09:55 2018
Thread 1 cannot allocate new log, sequence 8
Private strand flush not complete
Current log# 1 seq# 7 mem# 0: /u01/app/oracle/oradata/orcl/redo01.log
Thread 1 advanced to log sequence 8 (LGWR switch)
Current log# 2 seq# 8 mem# 0: /u01/app/oracle/oradata/orcl/redo02.log
Sat Feb 10 20:10:19 2018
Thread 1 cannot allocate new log, sequence 9
Private strand flush not complete
Current log# 2 seq# 8 mem# 0: /u01/app/oracle/oradata/orcl/redo02.log
Thread 1 advanced to log sequence 9 (LGWR switch)
Current log# 3 seq# 9 mem# 0: /u01/app/oracle/oradata/orcl/redo03.log
Thread 1 cannot allocate new log, sequence 10
Checkpoint not complete
Current log# 3 seq# 9 mem# 0: /u01/app/oracle/oradata/orcl/redo03.log
Thread 1 cannot allocate new log, sequence 10
Private strand flush not complete
Current log# 3 seq# 9 mem# 0: /u01/app/oracle/oradata/orcl/redo03.log
Thread 1 advanced to log sequence 10 (LGWR switch)
Current log# 1 seq# 10 mem# 0: /u01/app/oracle/oradata/orcl/redo01.log
这是为什么呢？那是因为其进程持有的句柄并有释放，如下
[oracle@togogo ~]$ ps -ef|grep ckpt|grep -v grep
oracle 8427 1 0 19:47 ? 00:00:00 ora_ckpt_orcl
[oracle@togogo ~]$ cd /proc/8427/fd
[oracle@togogo fd]$ ls -ltr |grep control
lrwx------ 1 oracle oinstall 64 Feb 10 20:09 257 -> /u01/app/oracle/fast_recovery_area/orcl/control02.ctl
lrwx------ 1 oracle oinstall 64 Feb 10 20:09 256 -> /u01/app/oracle/oradata/orcl/control01.ctl
#### session 1 trace跟踪
这里说一个命令 strace
举一个例子
strace -o output.txt -T -tt -e trace=all -p 28979
上面的含义是 跟踪28979进程的所有系统调用（-e trace=all），并统计系统调用的花费时间，以及开始时间（并以可视化的时分秒格式显示），最后将记录结果存在output.txt文件里面
[oracle@togogo fd]$ strace -fr -o /tmp/8427.log -p 8427
Process 8427 attached - interrupt to quit
####观察trace.log
8427 0.000156 gettimeofday({1518265363, 443228}, NULL) = 0
8427 0.000176 gettimeofday({1518265543, 106769}, NULL) = 0
8427 0.000077 gettimeofday({1518265543, 106845}, NULL) = 0
8427 0.000072 gettimeofday({1518265543, 106917}, NULL) = 0
8427 0.000077 pwrite64(256, "\25\302\0\0\3\0\0\0\0\0\0\0\0\0\1\4\312T\0\0\2\0\0\0\0\0\0\0Q\0\0\0"..., 16384, 49152) = 16384
8427 0.006462 gettimeofday({1518265543, 113463}, NULL) = 0
8427 0.000089 gettimeofday({1518265543, 113548}, NULL) = 0
8427 0.000080 pwrite64(257, "\25\302\0\0\3\0\0\0\0\0\0\0\0\0\1\4\312T\0\0\2\0\0\0\0\0\0\0Q\0\0\0"..., 16384, 49152) = 16384
8427 0.000734 gettimeofday({1518265543, 114364}, NULL) = 0
8427 0.000081 gettimeofday({1518265543, 114443}, NULL) = 0
8427 0.000081 gettimeofday({1518265543, 114525}, NULL) = 0
8427 0.000078 gettimeofday({1518265543, 114603}, NULL) = 0
8427 0.000211 gettimeofday({1518265543, 114816}, NULL) = 0
8427 0.000080 gettimeofday({1518265543, 114891}, NULL) = 0
8427 0.000081 pread64(256, "\25\302\0\0\1\0\0\0\0\0\0\0\0\0\1\4r\t\0\0\0\0\0\0\0\0 \v\373M\21Y"..., 16384, 16384) = 16384
这的 256 257 是表示的是文件描述符 FD file discription
16384是表示一个块的大小
49152是表示偏移量
通过以上可以得到结论
1. 进程信息可以在/proc下看到，例如： /proc/8427/stat
2. 对于linux，对于文件的读写，是通过调用函数read,pwrite64 来实现的。
3. 我们可以发现检查点进程ckpt 3s触发一次的机制。
4. 对于pwrite64的操作，是通过写fd （256.257）2个文件来完成的，其中对应的offset都是49152，且我们知道这3个被写入的文件 (我们知道是controlfile) 的block大小是16384.
