[20210224]控制文件序列号满的分析.txt
[20210224]控制文件序列号满的分析.txt
--//上午看了链接:https://blog.csdn.net/enmotech/article/details/113855641,出现控制文件序列号满的情况,我从来没有遇到.
--//下午没事,看看是否能在测试环境演示出来重复故障.
--//注意不能在生产系统做这样的测试!!!很久没有做这类恢复工作,写的有点乱.
1.环境:
scott@book> @ ver1
port_string version banner
------------------------------ -------------- --------------------------------------------------------------------------------
x86_64/linux 2.4.xx 11.2.0.4.0 oracle database 11g enterprise edition release 11.2.0.4.0 - 64bit production
2.测试:
--//首先我仔细看了原始链接,我发现作者通过设置小的闪回区,不断切换归档,看查看控制文件顺序号.
--//我自己也尝试了一下,导致归档满了,挂起,我认为跟这个没关系,实际上只要切换归档或者发出检查点,
--//控制文件顺序号就会增加.我先验证这种情况.
sqlplus -s -l scott/book << eof | grep "[0-9]"
set head off
set feedback off
$(seq 5 | xargs -i{} echo -e 'alter system checkpoint;\nselect controlfile_sequence# from v$database;\nhost sleep 1')
quit
eof
36500
36501
36502
36503
36504
--//你可以发现发出检查点,controlfile_sequence#增加1.
sqlplus -s -l scott/book << eof | grep "[0-9]"
set head off
set feedback off
$(seq 5 | xargs -i{} echo -e 'alter system switch logfile;\nselect controlfile_sequence# from v$database;\nhost sleep 1')
quit
eof
36506
36509
36513
36517
36520
--//执行alter system switch logfile也是一样.
--//我有一种预感可能对方日志可能切换过于频繁,可能导致控制文件序列号增加太快,消耗枯竭.
3.首先定位它在控制文件的什么位置.
sys@book> @ spid
sid serial# process server spid pid p_serial# c50
------------ ------------ ------------------------ --------- ------ ------- ------------ --------------------------------------------------
30 261 57742 dedicated 57743 26 101 alter system kill session '30,261' immediate;
scott@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
36571
$ strace -f -p 57743 -e pread
process 57743 attached - interrupt to quit
pread(256, "\25\302\0\0\1\0\0\0\0\0\0\0\0\0\1\4\327?\0\0\0\0\0\0\0\4 \vn!\267o"..., 16384, 16384) = 16384
pread(256, "\25\302\0\0\17\0\0\0j\275\0\0\377\377\1\4 c\0\0\0\6\0\0\0\0\0\0\0\0\0\4"..., 16384, 245760) = 16384
pread(256, "\25\302\0\0\21\0\0\0j\275\0\0\377\377\1\4\225z\0\0\0\0\0\0\0\0\0\0\256\36q5"..., 16384, 278528) = 16384
pread(256, "\25\302\0\0\1\0\0\0\0\0\0\0\0\0\1\4\327?\0\0\0\0\0\0\0\4 \vn!\267o"..., 16384, 16384) = 16384
pread(256, "\25\302\0\0\17\0\0\0j\275\0\0\377\377\1\4 c\0\0\0\6\0\0\0\0\0\0\0\0\0\4"..., 16384, 245760) = 16384
pread(256, "\25\302\0\0\21\0\0\0j\275\0\0\377\377\1\4\225z\0\0\0\0\0\0\0\0\0\0\256\36q5"..., 16384, 278528) = 16384
pread(256, "\25\302\0\0\32\1\0\0b\275\0\0\377\377\1\4\3270\0\0\27\2\2\0\345\7\0\0\0\0\0\0"..., 16384, 4620288) = 16384
--//应该读的位置在控制文件偏移16384,245760,278528,4620288.
$ echo 16384,245760,278528,4620288 | tr ',' '\n' | xargs -iq bash -c "echo q/16384| bc "| paste -sd','
1,15,17,282
--//关闭数据库,重新启动数据库到mount状态.这样启动后应该不变.
sys@book> startup mount
oracle instance started.
total system global area 643084288 bytes
fixed size 2255872 bytes
variable size 205521920 bytes
database buffers 427819008 bytes
redo buffers 7487488 bytes
database mounted.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
36580
--//36580 = 0x8ee4,颠倒过来就是0xe48e.
$ echo 1,15,17,282 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x e48e top' | rlbbed
bbed: release 2.0.0.0.0 - limited production on wed feb 24 15:59:39 2021
copyright (c) 1982, 2011, oracle and/or its affiliates. all rights reserved.
************* !!! for oracle internal use only !!! ***************
bbed> set count 64
count 64
bbed> set width 160
width 160
bbed> dba 0x19400001 (423624705 101,1)
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 103 dba:0x19400001
------------------------------------------------------------------------------------------------------------------------------------------------
e48e0000 8c020000 00400000 00000100 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 5e89c659 ae1e7135
<64 bytes per line>
bbed> dba 0x1940000f (423624719 101,15)
bbed> bbed-00212: search string not found
bbed> dba 0x19400011 (423624721 101,17)
bbed> bbed-00212: search string not found
bbed> dba 0x1940011a (423624986 101,282)
bbed> bbed-00212: search string not found
--//注我已经定义101对应控制文件,并且实际上指定101,自动设置blocksize=16384,不需要设定.
$ ls -l /mnt/ramdisk/book/control01.ctl
-rw-r----- 1 oracle oinstall 10698752 2021-02-24 16:03:22 /mnt/ramdisk/book/control01.ctl
--//10698752/16384 = 653,从0开始到652,控制文件的0块也是os头,你可以使用xxd -c 16 /mnt/ramdisk/book/control01.ctl查看.
$ xxd -c 16 /mnt/ramdisk/book/control01.ctl | head -6
0000000: 00c2 0000 0000 c0ff 0000 0000 0000 0000 .?...?........
0000010: eaf8 0000 0040 0000 8c02 0000 7d7c 7b7a 犋...@......}||z
~~~~~~~~~=>奇幻数,数据文件以及日志文件都有类似标识.
0000020: a081 0000 0000 0000 0000 0000 0000 0000 ................
0000030: 0000 0000 0000 0000 0000 0000 0000 0000 ................
0000040: 0000 0000 0000 0000 0000 0000 0000 0000 ................
0000050: 0000 0000 0000 0000 0000 0000 0000 0000 ................
$ seq 652 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x e48e top' | rlbbed | grep -c2 "\-\-\--"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 103 dba:0x19400001
------------------------------------------------------------------------------------------------------------------------------------------------
e48e0000 8c020000 00400000 00000100 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 5e89c659 ae1e7135
--//很明显位置在控制文件块1,偏移40的位置,应该占32位,4个字节.最大0xffffffff = 4294967295.
4.先做一些功课,看看数据文件以及日志是否存在这个信息.
--//我看一些文件头资料,在数据文件头kcvfh.kcvfhhdr.kccfhcsq也是记录控制文件的seq.
$ seq 6 | xargs -iq echo "p dba q,1 kcvfh.kcvfhhdr.kccfhcsq" | rlbbed| grep kccfhcsq
bbed> ub4 kccfhcsq @40 0x00008edc
bbed> ub4 kccfhcsq @40 0x00008edc
bbed> ub4 kccfhcsq @40 0x00008edc
bbed> ub4 kccfhcsq @40 0x00008edc
bbed> ub4 kccfhcsq @40 0x00008edc
bbed> ub4 kccfhcsq @40 0x00008edc
--//也是在文件头偏移40的位置,但是数值0x8edc = 36572.感觉是我关闭数据库记录的值写入的,比前面+1(对比前面).
--//看看日志文件的情况:
$ seq 501 503 | xargs -iq echo "dump /v dba q,1 offset 36 count 8" | rlbbed | grep 00900100
d68e0000 00900100 l ........
dc8e0000 00900100 l ........
d28e0000 00900100 l ........
--//注我已经定义501,502,503对应日志文件redo01.log,redo02.log,redo03.log.
--//我估计在切换日志时写入当时的控制文件seq.
sys@book> @ log
show redo log layout from v$log and v$logfile...
group# thread# sequence# bytes blocksize members arc status first_change# first_time next_change# next_time
--------------- --------------- --------------- --------------- --------------- --------------- --- ---------------- --------------- ------------------- --------------- -------------------
1 1 716 52428800 512 1 yes inactive 13276955579 2021-02-24 15:48:54 13276955587 2021-02-24 15:48:57
2 1 717 52428800 512 1 no current 13276955587 2021-02-24 15:48:57 281474976710655
3 1 715 52428800 512 1 yes inactive 13276955571 2021-02-24 15:48:51 13276955579 2021-02-24 15:48:54
--//可以确定日志文件也有对应信息.位于块1的偏移36字节处.当前日志是第2组.
5.尝试修改控制文件顺序号:
--//注意修改前备份控制文件.关闭数据库
$ cp /mnt/ramdisk/book/control0* /home/oracle/tmp/
bbed> set dba 101,1
dba 0x19400001 (423624705 101,1)
bbed> dump /v dba 101,1 offset 40 count 8
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 47 dba:0x19400001
-----------------------------------------------------------------------------------------------------------
e68e0000 8c020000 l ........
<32 bytes per line>
--//关闭数据库后增加+2.
$ seq 652 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x e68e top' | rlbbed | grep -c2 "\-\-\--"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 103 dba:0x19400001
------------------------------------------------------------------------------------------------------------------------------------------------
e68e0000 8c020000 00400000 00000100 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 5e89c659 ae1e7135
$ seq 652 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x e48e top' | rlbbed | grep -c2 "\-\-\--"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 15 offsets: 8 to 71 dba:0x1940000f
------------------------------------------------------------------------------------------------------------------------------------------------
e48e0000 ffff0104 2ce80000 00040000 00000000 00000004 06440008 000400e3 00000000 00000002 00000000 00000000 008ddae0 15000000 00000000 00000000
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 17 offsets: 8 to 71 dba:0x19400011
------------------------------------------------------------------------------------------------------------------------------------------------
e48e0000 ffff0104 10de0000 00000000 00000000 ae1e7135 424f4f4b 00000000 00000000 08024000 01404010 00000000 00000000 06200e00 00000000 b01e7135
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 281 offsets: 8 to 71 dba:0x19400119
------------------------------------------------------------------------------------------------------------------------------------------------
e48e0000 ffff0104 9adc0000 0dff0200 e1070000 00000000 00000000 02000000 02000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
--//视乎在别的位置备份原来的控制文件seq信息.仅仅是猜测.因为前面查询没找到.
--//我尝试再次mount查询该位置.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
36587
--//36587 = 0x8eeb
$ seq 652 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x eb8e top' | rlbbed | grep -c2 "\-\-\--"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 103 dba:0x19400001
------------------------------------------------------------------------------------------------------------------------------------------------
eb8e0000 8c020000 00400000 00000100 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 12d1c659 ae1e7135
$ seq 652 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\nfind /x ea8e top' | rlbbed | grep -c2 "\-\-\--"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 15 offsets: 8 to 71 dba:0x1940000f
------------------------------------------------------------------------------------------------------------------------------------------------
ea8e0000 ffff0104 2ca00000 00040000 00000000 00000004 06440008 000c00a3 04000000 00000002 00000000 04000000 008ddae0 15000000 00000000 00000000
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 17 offsets: 8 to 71 dba:0x19400011
------------------------------------------------------------------------------------------------------------------------------------------------
ea8e0000 ffff0104 10de0000 00000000 00000000 ae1e7135 424f4f4b 00000000 00000000 08024000 01404010 00000000 00000000 06200e00 00000000 b01e7135
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 319 offsets: 8 to 71 dba:0x1940013f
------------------------------------------------------------------------------------------------------------------------------------------------
ea8e0000 ffff0104 e60b0000 626f6f6b 000060ff ffffffff 03b66b2a ff7f0000 38aeef7e 00000000 48000000 00000000 48000000 fd7f0000 10000000 00000000
--//应该可以肯定我的分析大致正确.开始修改看看.注意最好在关闭数据库下进行.
bbed> dump /v dba 101,1 offset 40 count 16
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 55 dba:0x19400001
-----------------------------------------------------------------------------------------------------------
ed8e0000 8c020000 00400000 00000100 l .........@......
<32 bytes per line>
--//修改命令如下:
modify /x ff8e dba 101,1 offset 40
sum apply dba 101,1
modify /x ff8e dba 102,1 offset 40
sum apply dba 102,1
--//注意另外的控制文件也要修改:
bbed> modify /x ff8e dba 101,1 offset 40
warning: contents of previous bifile will be lost. proceed? (y/n) y
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 55 dba:0x19400001
------------------------------------------------------------------------------------------------------------------------------------------------
ff8e0000 8c020000 00400000 00000100
<64 bytes per line>
bbed> sum apply dba 101,1
check value for file 101, block 1:
current = 0x8883, required = 0x8883
bbed> modify /x ff8e dba 102,1 offset 40
file: /mnt/ramdisk/book/control02.ctl (102)
block: 1 offsets: 40 to 55 dba:0x19800001
------------------------------------------------------------------------------------------------------------------------------------------------
ff8e0000 8c020000 00400000 00000100
<64 bytes per line>
bbed> sum apply dba 102,1
check value for file 102, block 1:
current = 0x8883, required = 0x8883
--//0x8eff = 36607,验证看看:
sys@book> startup mount
oracle instance started.
total system global area 643084288 bytes
fixed size 2255872 bytes
variable size 205521920 bytes
database buffers 427819008 bytes
redo buffers 7487488 bytes
database mounted.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
36612
--//36612 = 0x8f04,即使在mount状态也增加一点点.
bbed> dump /v dba 101,1 offset 40 count 16
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 55 dba:0x19400001
-----------------------------------------------------------------------------------------------------------
048f0000 8c020000 00400000 00000100 l .........@......
<32 bytes per line>
--//说明修改位置正确.
6.继续增大步幅:
--//注意修改前最好关闭数据库进行.
bbed> modify /x 8affffff dba 101,1 offset 40
bbed-00209: invalid number (8affffff)
--//有一个小技巧就是如果修改4个字节的话,第一个字符要小于0x8,,不然报错.
--//修改命令如下:
modify /x 7fffffff dba 101,1 offset 40
sum apply dba 101,1
modify /x 7fffffff dba 102,1 offset 40
sum apply dba 102,1
--//相当于控制文件seq = 0xffffff7f = 4294967167.
--//0xff-0x7f = 128
sys@book> startup mount
oracle instance started.
total system global area 643084288 bytes
fixed size 2255872 bytes
variable size 205521920 bytes
database buffers 427819008 bytes
redo buffers 7487488 bytes
database mounted.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
4294967172
--//4294967172= 0xffffff84,^_^现在已经逼近消耗枯竭状态了.看看是否可以打开,该是见证奇迹的时刻...
sys@book> alter database open ;
database altered.
sys@book> alter system checkpoint ;
system altered.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
4294967181
sys@book> alter system checkpoint ;
system altered.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
4294967182
--//每次checkpoint仅仅增加1. 0xffffffff = 4294967295
sys@book> alter system switch logfile;
system altered.
--//执行多次,加快增加....省略....
sys@book> alter system switch logfile;
system altered.
..
sys@book> alter system checkpoint ;
system altered.
sys@book> select controlfile_sequence# from v$database;
controlfile_sequence#
---------------------
4294967295
--//4294967295= 0xffffffff,在往前呢?
sys@book> alter system checkpoint ;
system altered.
sys@book> select controlfile_sequence# from v$database;
--//挂起..
--//alert.log报如下错误.
wed feb 24 17:14:01 2021
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_arc1_29847.trc:
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_arc1_29847.trc (incident=3792203):
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
incident details in: /u01/app/oracle/diag/rdbms/book/book/incident/incdir_3792203/book_arc1_29847_i3792203.trc
wed feb 24 17:14:02 2021
sweep [inc][3792203]: completed
sweep [inc][3792195]: completed
sweep [inc2][3792203]: completed
sweep [inc2][3792195]: completed
dumping diagnostic data in directory=[cdmp_20210224171402], requested by (instance=1, osid=29847 (arc1)), summary=[incident=3792203].
bbed> dump /v dba 101,1 offset 40 count 16
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 55 dba:0x19400001
-----------------------------------------------------------------------------------------------------------
00000000 8c020000 00400000 00000100 l .........@......
<32 bytes per line>
--//全部为0.补充说明我做了多次,如果seq跳跃很快,数据库直接crash,我上面的测试到4294967295时换成了alter system checkpoint ;.
$ echo 15,17,281,319 | tr ',' '\n' | xargs -iq echo -e 'set dba 101,q\ndump /v offset 8 count 8' | rlbbed | grep -c2 "\-\-"
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 15 offsets: 8 to 15 dba:0x1940000f
-----------------------------------------------------------------------------------------------------------
ffffffff ffff0104 l ........
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 17 offsets: 8 to 15 dba:0x19400011
-----------------------------------------------------------------------------------------------------------
ffffffff ffff0104 l ........
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 281 offsets: 8 to 15 dba:0x19400119
-----------------------------------------------------------------------------------------------------------
85ffffff ffff0104 l ........
--
bbed> file: /mnt/ramdisk/book/control01.ctl (101)
block: 319 offsets: 8 to 15 dba:0x1940013f
-----------------------------------------------------------------------------------------------------------
83ffffff ffff0104 l ........
--//验证我的推测.这里也记录seq信息.
$ seq 501 503 | xargs -iq echo "dump /v dba q,1 offset 36 count 8" | rlbbed | grep 00900100
f8ffffff 00900100 l ........
fcffffff 00900100 l ........
fcffffff 00900100 l ........
--//现在数据库依旧可以登录,但是一些查询会hang住,比如查询v$database视图.
sys@book> shutdown immediate;
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
sys@book> alter system checkpoint ;
alter system checkpoint
*
error at line 1:
ora-03113: end-of-file on communication channel
process id: 30040
session id: 114 serial number: 3
--//alert.log记录如下:
wed feb 24 17:27:02 2021
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_ckpt_29816.trc:
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_ckpt_29816.trc (incident=3792131):
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_ckpt_29816.trc:
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
ckpt (ospid: 29816): terminating the instance due to error 227
wed feb 24 17:27:03 2021
system state dump requested by (instance=1, osid=29816 (ckpt)), summary=[abnormal instance termination].
system state dumped to trace file /u01/app/oracle/diag/rdbms/book/book/trace/book_diag_29800_20210224172703.trc
dumping diagnostic data in directory=[cdmp_20210224172703], requested by (instance=1, osid=29816 (ckpt)), summary=[abnormal instance termination].
instance terminated by ckpt, pid = 29816
7.尝试恢复:
sys@book> startup mount
oracle instance started.
total system global area 643084288 bytes
fixed size 2255872 bytes
variable size 205521920 bytes
database buffers 427819008 bytes
redo buffers 7487488 bytes
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
sys@book> select controlfile_sequence# from v$database;
select controlfile_sequence# from v$database
*
error at line 1:
ora-01507: database not mounted
--//alert.log
alter database mount
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_ora_30094.trc:
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
wed feb 24 17:29:02 2021
sweep [inc][3792131]: completed
errors in file /u01/app/oracle/diag/rdbms/book/book/trace/book_ora_30094.trc (incident=3793788):
ora-00227: corrupt block detected in control file: (block 1, # blocks 1)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-// 我没有遇到原链接的提示,这个提示有点昏.
ora-00202: control file: '/mnt/ramdisk/book/control01.ctl'
incident details in: /u01/app/oracle/diag/rdbms/book/book/incident/incdir_3793788/book_ora_30094_i3793788.trc
dumping diagnostic data in directory=[cdmp_20210224172902], requested by (instance=1, osid=30094), summary=[incident=3793788].
ora-227 signalled during: alter database mount...
$ dbv file=/mnt/ramdisk/book/control01.ctl blocksize=16384
dbverify: release 11.2.0.4.0 - production on wed feb 24 17:29:52 2021
copyright (c) 1982, 2011, oracle and/or its affiliates. all rights reserved.
dbverify - verification starting : file = /mnt/ramdisk/book/control01.ctl
dbverify - verification complete
total pages examined : 652
total pages processed (data) : 0
total pages failing (data) : 0
total pages processed (index): 0
total pages failing (index): 0
total pages processed (other): 147
total pages processed (seg) : 0
total pages failing (seg) : 0
total pages empty : 505
total pages marked corrupt : 0
total pages influx : 0
total pages encrypted : 0
highest block scn : 4294967295 (65535.4294967295)
--//ok,但是注意后面那行.highest block scn : 4294967295 (65535.4294967295)
--//感觉这个可以称为控制文件的scn号. 4294967295 = 0xffffffff ,65535 = 0xffff
$ seq 6 | xargs -iq echo "p dba q,1 kcvfh.kcvfhhdr.kccfhcsq" | rlbbed| grep kccfhcsq
bbed> ub4 kccfhcsq @40 0xffffffff
bbed> ub4 kccfhcsq @40 0xffffffff
bbed> ub4 kccfhcsq @40 0xffffffff
bbed> ub4 kccfhcsq @40 0xffffffff
bbed> ub4 kccfhcsq @40 0xffffffff
bbed> ub4 kccfhcsq @40 0xffffffff
--//数据文件的kccfhcsq全部是0xffffffff.
bbed> dump /v dba 101,1 offset 40 count 4
file: /mnt/ramdisk/book/control01.ctl (101)
block: 1 offsets: 40 to 43 dba:0x19400001
-----------------------------------------------------------------------------------------------------------
00000000 l ....
<32 bytes per line>
--//执行如下:
modify /x 7fffffff dba 101,1 offset 40
modify /x 7fffffff dba 102,1 offset 40
modify /x ff dba 101,1 offset 40
modify /x ff dba 102,1 offset 40
sum apply dba 101,1
sum apply dba 102,1
--//重启不行.
modify /x 7fffffff dba 101,1 offset 40
modify /x 7fffffff dba 102,1 offset 40
sum apply dba 101,1
sum apply dba 102,1
sys@book> startup mount
oracle instance started.
total system global area 643084288 bytes
fixed size 2255872 bytes
variable size 205521920 bytes
database buffers 427819008 bytes
redo buffers 7487488 bytes
ora-00600: internal error code, arguments: [2131], [9], [8], [], [], [], [], [], [], [], [], []
--//不行.
7.如何恢复呢?
--//链接介绍一种方式就是重建控制文件,使用resetlogs重建.
--//原始链接的测试使用resetlogs重建的控制文件,我做了ok,不再贴出,方法比较简单.
--//我想给自己增加一点点难度,就是使用noresetlogs打开,因为这样重建的控制文件要读取redo,数据文件重新
--//回填一些信息,实际上resetlogs也类似,但是noresetlogs回填的控制文件seq很大,一样打不开数据库.
--//也就是必须提到我前面要修改的数据文件以及redo文件的几个偏移位置.太长了,另外写一篇blog.
8.补充说明:
--//前面我提到的rlbbed我建立的bash shell函数,你可以使用别名代替,效果一样的.
$ export rlwrap=$(which rlwrap)
$ type rlbbed
rlbbed is a function
rlbbed ()
{
cd /home/oracle/bbed;
$rlwrap -s 9999 -c -r -i -f /usr/local/share/rlwrap/bbed $oracle_home/bin/bbed parfile=bbed.par cmdfile=cmd.par
}
--//关于bbed配置看相关文档.
$ cat cmd.par
set count 64
set width 160
$ cat bbed.par
blocksize=8192
listfile=$home/bbed/filelist.txt
mode=edit
password=blockedit
spool=y
--//filelist.txt文件通过select file#||' '||name c100 from v$dbfile order by file#;生成.我增加了控制文件以及redo文件.
--//还有临时文件.
$ cat filelist.txt | grep -v "#"
4 /mnt/ramdisk/book/users01.dbf
1 /mnt/ramdisk/book/system01.dbf
2 /mnt/ramdisk/book/sysaux01.dbf
3 /mnt/ramdisk/book/undotbs01.dbf
5 /mnt/ramdisk/book/example01.dbf
6 /mnt/ramdisk/book/tea01.dbf
101 /mnt/ramdisk/book/control01.ctl
102 /mnt/ramdisk/book/control02.ctl
201 /mnt/ramdisk/book/temp01.dbf
501 /mnt/ramdisk/book/redo01.log
502 /mnt/ramdisk/book/redo02.log
503 /mnt/ramdisk/book/redo03.log
504 /mnt/ramdisk/book/redostb01.log
上一篇： Linux命令unzip详解
下一篇： 深入浅出分析C++ string底层原理