728x90

인포믹스 데이터 파일 내용을 조작하여 백업 테스트를 수행해보았습니다.


## 테스트 환경

1. 테스트서버 : AIX 6.1

2. DBSPACE 현황 (청크개수) : rootdbs (1), datadbs (3)

3. onstat -d 결과

IBM Informix Dynamic Server Version 7.31.UD9     -- Quiescent -- Up 00:29:45 -- 18112 Kbytes


Dbspaces

address  number   flags    fchunk   nchunks  flags    owner    name

4005a150 1        1        1        1        N        informix rootdbs

4005a888 2        1        2        1        N        informix plogdbs

4005a948 3        1        3        1        N        informix llogdbs

4005aa08 4        1        4        3        N        informix datadbs

 4 active, 2047 maximum


Chunks

address  chk/dbs offset   size     free     bpages   flags pathname

4005a210 1   1   0        5000     3009              PO-   /work2/INFORMIX/ids731/dbspaces/rootdbs

4005a388 2   2   0        250000   249947            PO-   /work2/INFORMIX/ids731/dbspaces/plogdbs

4005a488 3   3   0        250000   249707            PO-   /work2/INFORMIX/ids731/dbspaces/llogdbs

4005a588 4   4   0        250000   0                 PO-   /work2/INFORMIX/ids731/dbspaces/datadbs

4005a688 5   4   0        25000    0                 PO-   /work2/INFORMIX/ids731/dbspaces/datadbs1

4005a788 6   4   0        25000    0                 PO-   /work2/INFORMIX/ids731/dbspaces/datadbs2

 6 active, 2047 maximum


 

 

## 테스트 절차

1. 데이터 파일 조작 (6 청크 파일)

dd if=/dev/random seek=10 bs=1000 count=10 of=/work2/INFORMIX/ids731/dbspaces/datadbs2 conv=notrunc

 

2. ontape 백업

ontape -s -L 0

 

3. 백업 online.log 파일 내용 확인

17:37:39  Assert Failed: Archive detects that page 0x600003 is corrupt.

17:37:39  IBM Informix Dynamic Server Version 7.31.UD9

17:37:39   Who: Session(23, informix@pilma01, 23134372, 1074292452)

                Thread(60, arcbackup1, 4005e334, 1)

                File: rsarcbu.c Line: 2600

17:37:39  stack trace for pid 23396402 written to /tmp/af.424e052

17:37:39   See Also: /tmp/af.424e052

17:37:39  Archive on rootdbs, datadbs Completed with 2 corrupted pages detected.


백업중 발생한 오류 메시지에서 나타난 페이지 번호(0x600003)는 DBSPace 번호가 아닌 Chunk 번호를 의미하는 것으로 보입니다.

728x90
728x90

The message Archive detects that page is corrupt indicates that page validation failed.

During an archive, the database server validates every page before writing it to the archive device. This validation checks that the elements on the page are consistent with the expected values. When a page fails this validation, a message similar to the following is written to the online.log file:

15:06:37  Assert Failed: Archive detects that page 0xc00021 is corrupt.
15:06:37  IBM Informix Dynamic Server Version 9.40.UC1
15:06:37   Who: Session(25, informix@cronus, 67612, 1085259772)
                Thread(50, arcbackup1, 40acf758, 4)
                File: rsarcbu.c Line: 2549
15:06:37  stack trace for pid 67367 written to /tmp/af.41ad7b9
15:06:37   See Also: /tmp/af.41ad7b9

below 9.40 version (Expanded chunk capacity mode: disabled)
16:27:49 Assert Warning: Archive detects that page 1:10164 is corrupt. 
16:27:49 IBM Informix Dynamic Server Version 11.50.FC7
16:27:49 Who: Session(5, informix@cronus, 23467, 10a921048)
Thread(40, arcbackup1, 10a8e8ae8, 1)
File: rsarcbu.c Line: 2915
16:27:49 stack trace for pid 23358 written to /tmp/af.41043f4
16:27:49 See Also: /tmp/af.41043f4
16:27:49 Archive detects that page 1:10164 is corrupt.
16:27:50 Archive on rootdbs Completed with 1 corrupted pages detected.

above 10.00 version (Expanded chunk capacity mode: enabled, always)

The page number is printed in hexadecimal. The format for page number is 0xCCCPPPPP where CCC represents the chunk number, and PPPPP represents the page number. For this example, the corrupted page is in chunk 0xc (12 decimal) and page 0x21 (33 decimal). The archive aborts after detecting 10 corrupt pages. The online.log file displays the full error message, including the page address, for the first 10 errors. Subsequently, only the count of the number of corrupt pages is put in to the online.log.

When you receive this message, identify which table the corrupt page belongs to by examining the output of the oncheck –pe command. (해당 정보는 sysmaster:sysextents 테이블에서도 확인할 수 있습니다.) To determine the extent of the corruption, execute the oncheck –cID command for that table.

A corrupt page is saved onto the backup media. During a restore, the corrupt page is returned in its corrupt form. No errors messages are written to the online.log when corrupt pages are restored, only when they are archived.


http://publib.boulder.ibm.com/infocenter/idshelp/v10/index.jsp?topic=/com.ibm.bar.doc/barmst369.htm

http://publib.boulder.ibm.com/infocenter/idshelp/v115/index.jsp?topic=%2Fcom.ibm.bar.doc%2Fids_bar_055.htm

http://www-01.ibm.com/support/docview.wss?uid=swg21209842


728x90

+ Recent posts