通过trace文件,跟踪数据库的启动过程
[@more@]将数据库启动到nomount阶段,打开10046事件对启动过程进行跟踪。
sql> startup nomount
sql> alter session set events '10046 trace name context forever, level 12';
sql> alter database mount;
---trace文件中相应的记录
WAIT #3: nam='control file sequential read' ela= 1863 file#=0 block#=3 blocks=8 obj#=-1 tim=1251077602436218
WAIT #3: nam='control file sequential read' ela= 252 file#=1 block#=3 blocks=8 obj#=-1 tim=1251077602436503
读取控制文件记录。
sql> alter database open;
---trace文件中相应的记录
WAIT #1: nam='direct path read' ela= 608 file number=1 first dba=1 block cnt=1 obj#=-1 tim=1251077621080498
WAIT #1: nam='direct path read' ela= 2 file number=2 first dba=1 block cnt=1 obj#=-1 tim=1251077621080551
通过direct path read,直接读取数据文件的第一个块(文件头),获取scn,checkpoint等等信息;
WAIT #1: nam='db file sequential read' ela= 59 file#=1 block#=377 blocks=1 obj#=-1 tim=1251077621378261
create table bootstrap$ ( line# number not null, obj# number not null, sql_text varchar2(4000) not null) s
torage (initial 50K objno 56 extents (file 1 block 377))
通过'db file sequential read' 读取system表空间拥有的数据文件的377号块,并以此创建bootstrap$表,开启bootstrap引导过程。
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/56630/viewspace-1036368/,如需转载,请注明出处,否则将追究法律责任。
转载于:http://blog.itpub.net/56630/viewspace-1036368/