[20190124]bbed恢复数据遇到延迟块清除的问题.txt

[20190124]bbed恢复数据遇到延迟块清除的问题.txt --//最近使用bbed做一个恢复测试,遇到一个问题.以前我的测试如果修改删除flag从0x3c=>0x2c,sum apply后,使用verify提示类似如下: BBED> verify DBVERIFY - Verification starting FILE = /mnt/ramdisk/book/users01.dbf BLOCK = 523 Block Checking: DBA = 16777739, Block Type = KTB-managed data block data header at 0x7fddbce4127c kdbchk: the amount of space used is not equal to block size         used=44 fsc=9 avsp=8020 dtl=8064 Block 523 failed with check code 6110 --//如果偷懒,可以跳过这步.但是如果遇到提交时数据块不在缓存或者更新涉及的块太多,可能会出现许多块不做块清除,oracle执行的是 --//快速块清除操作.这样一些块在下一次touch时才修改对应ITL操以及对应记录的lock信息才会更新. --//对于这样的块,恢复时恢复会遇到什么问题呢?通过例子说明问题. 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.建立测试环境: SCOTT@book> create table t as select rownum id,'test' name from dual connect by level<=2; Table created. SCOTT@book> select rowid,t.* from t; ROWID                      ID NAME ------------------ ---------- -------------------- AAAWPYAAEAAAAILAAA          1 test AAAWPYAAEAAAAILAAB          2 test SCOTT@book> @ rowid AAAWPYAAEAAAAILAAA     OBJECT       FILE      BLOCK        ROW ROWID_DBA            DBA                  TEXT ---------- ---------- ---------- ---------- -------------------- -------------------- ----------------------------------------      91096          4        523          0  0x100020B           4,523                alter system dump datafile 4 block 523 ; SCOTT@book> delete from t where id=1; 1 row deleted. SCOTT@book> alter system flush buffer_cache; System altered. SCOTT@book> alter system flush buffer_cache; System altered. --//注:一定要确定对于数据块信息已经刷新到磁盘.IMU环境有时候奇怪,脏块有时候也不会写盘. SYS@book> @ bh 4 523 HLADDR           DBARFIL DBABLK CLASS CLASS_TYPE STATE TCH CR_SCN_BAS CR_SCN_WRP CR_UBA_FIL CR_UBA_BLK CR_UBA_SEQ BA               OBJECT_NAME ---------------- ------- ------ ----- ---------- ----- --- ---------- ---------- ---------- ---------- ---------- ---------------- ----------- 000000008538CA18       4    523     1 data block free    0          0          0          0          0          0 0000000073D1C000 T 000000008538CA18       4    523     1 data block free    0          0          0          0          0          0 0000000073D22000 T SCOTT@book> commit ; Commit complete. 3.使用bbed恢复看看: BBED> set dba 4,523         DBA             0x0100020b (16777739 4,523) BBED> x /rnc *kdbr[0] rowdata[11]                                 @8177 ----------- flag@8177: 0x3c (KDRHFL, KDRHFF, KDRHFD, KDRHFH) lock@8178: 0x02 cols@8179:    0 --//使用ITL槽2.看看ITL槽2(从0开始)的情况: BBED> p ktbbh.ktbbhitl[1] struct ktbbhitl[1], 24 bytes                @68    struct ktbitxid, 8 bytes                 @68       ub2 kxidusn                           @68       0x000a       ub2 kxidslt                           @70       0x0013       ub4 kxidsqn                           @72       0x00005816    struct ktbituba, 8 bytes                 @76       ub4 kubadba                           @76       0x00c0028e       ub2 kubaseq                           @80       0x10d5       ub1 kubarec                           @82       0x0f    ub2 ktbitflg                             @84       0x0001 (NONE)    union _ktbitun, 2 bytes                  @86       sb2 _ktbitfsc                         @86       9       ub2 _ktbitwrp                         @86       0x0009    ub4 ktbitbas                             @88       0x00000000 --//可以发现ktbitflg=0x0001,表示没有提交.ktbitbas=0x00000000,也就是没有scn相关信息写入. BBED> assign offset 8177=0x2c; Warning: contents of previous BIFILE will be lost. Proceed? (Y/N) y ub1 rowdata[0]                              @8177     0x2c BBED> x /rnc *kdbr[0] rowdata[11]                                 @8177 ----------- flag@8177: 0x2c (KDRHFL, KDRHFF, KDRHFH) lock@8178: 0x02 cols@8179:    2 col    0[2] @8180: 1 col    1[4] @8183: Test BBED> sum apply Check value for File 4, Block 523: current = 0xdcbe, required = 0xdcbe BBED> verify DBVERIFY - Verification starting FILE = /mnt/ramdisk/book/users01.dbf BLOCK = 523 Block Checking: DBA = 16777739, Block Type = KTB-managed data block data header at 0x7fddbce4127c kdbchk: the amount of space used is not equal to block size         used=44 fsc=9 avsp=8020 dtl=8064 Block 523 failed with check code 6110 --//我以前测试提到过这样恢复,读取是没有问题,虽然verify时包如上的错误. SCOTT@book> alter system flush buffer_cache; System altered. SCOTT@book> select rowid,t.* from t; ROWID                      ID NAME ------------------ ---------- -------------------- AAAWPZAAEAAAAILAAA          1 Test AAAWPZAAEAAAAILAAB          2 Test SCOTT@book> alter system flush buffer_cache; System altered. BBED> x /rnc dba 4,523 *kdbr[0] rowdata[11]                                 @8177 ----------- flag@8177: 0x2c (KDRHFL, KDRHFF, KDRHFH) lock@8178: 0x00 cols@8179:    2 col    0[2] @8180: 1 col    1[4] @8183: Test BBED> p dba 4,523 ktbbh.ktbbhitl[1] struct ktbbhitl[1], 24 bytes                @68    struct ktbitxid, 8 bytes                 @68       ub2 kxidusn                           @68       0x000a       ub2 kxidslt                           @70       0x0013       ub4 kxidsqn                           @72       0x00005816    struct ktbituba, 8 bytes                 @76       ub4 kubadba                           @76       0x00c0028e       ub2 kubaseq                           @80       0x10d5       ub1 kubarec                           @82       0x0f    ub2 ktbitflg                             @84       0xa000 (KTBFUPB, KTBFCOM)    union _ktbitun, 2 bytes                  @86       sb2 _ktbitfsc                         @86       3       ub2 _ktbitwrp                         @86       0x0003    ub4 ktbitbas                             @88       0x1774676a --//OK,可以发现没有问题.即使现在还报错: BBED> verify dba 4,523 DBVERIFY - Verification starting FILE = /mnt/ramdisk/book/users01.dbf BLOCK = 523 Block Checking: DBA = 16777739, Block Type = KTB-managed data block data header at 0x13efe7c kdbchk: the amount of space used is not equal to block size         used=44 fsc=0 avsp=8029 dtl=8064 Block 523 failed with check code 6110 --//如果使用system表空间呢?下一篇测试看看.