BAD INDEX occuring repeatedly...

[Former Member]

Hello,

we are experiencing repeated and increasing BAD INDEX Problems in one of our MaxDB instances. So far we have not been able to identify the cause. The database did not crash, but due to the enormous performance loss, the database literally "locked up" with connections failing etc.

Is there anything we can do to avoid these problems ??

Greetings,

Arne Gehlhaar

I have attached extracts from th knldiag and knldiag.err files.

2007-11-19 08:33:08 1 12781 INFO Version: 'Linux thread implementation is more POSIX compliant (NPTL)'

2007-11-19 08:33:08 1 12781 INFO Version: 'Linux thread implementation allows setting of individual thread stack size'

2007-11-19 08:33:08 2 12821 TASKING Thread 2 starting

2007-11-19 08:33:08 1 12600 VERSION 'Kernel 7.6.00 Build 016-123-109-428'

2007-11-19 08:33:08 1 12600 VERSION 'X64/LINUX 7.6.00 Build 016-123-109-428'

2007-11-19 08:33:08 1 12769 INFO Starting SERVERDB: 'CODABA'

2007-11-19 08:33:08 1 12770 INFO SERVERNODE: 'kundenserver'

2007-11-19 08:33:08 1 12773 INFO Date: 2007-11-19

2007-11-19 08:33:08 1 12771 INFO Process ID: 17582

2007-11-19 08:33:08 1 12772 INFO Owner: 'sdb'

2007-11-19 08:33:08 1 12772 INFO Group: 'sdba'

2007-11-19 08:33:08 1 12781 INFO Version: '64BIT Kernel'

2007-11-19 08:33:08 1 12781 INFO Version: 'X64/LINUX 7.6.00 Build 016-123-109-428'

2007-11-19 08:33:08 1 12781 INFO Version: 'FAST'

2007-11-19 08:33:08 1 12806 INFO Machine: 'x86_64'

2007-11-19 08:33:08 1 12780 INFO DBROOT: '/var/opt/sdb/data/wrk/CODABA'

2007-11-19 08:33:08 1 12780 INFO LD_LIBRARY_PATH: '/opt/sdb/7600/lib:/opt/sdb/7600/sap'

knldiag.err

2007-11-22 05:32:26 143 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

2007-11-22 05:32:26 143 ERR 20011 MOVECODE VBD600:1

2007-11-22 05:32:26 143 ERR 20011 MOVECODE 2007-11-22 05:32:26 MOVECODE Error 20011

2007-11-22 05:32:26 143 ERR 20011 MOVECODE Bad parameter: limit source 8181 destination 8181, source [0X2AA43FE000]7863, destination [0X2AA43FE000]7873, -19

2007-11-22 05:32:26 143 ERR 20011 MOVECODE 28 bytes

2007-11-22 05:32:26 143 ERR 53000 B*TREE 07010000000000018EB7000000000000

2007-11-22 05:32:26 143 ERR 53000 B*TREE Index Root 939069

2007-11-22 05:32:26 143 ERR 53333 B*TREE Data page corrupted: 865464

2007-11-22 05:32:26 143 ERR 53250 INDEX Bad Index 939069 (Root)

2007-11-22 05:32:26 143 ERR 53250 INDEX Reason "System error: BD Data page corr"

2007-11-22 05:32:26 143 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

2007-11-22 05:32:26 143 ERR 20011 MOVECODE VBD600:1

2007-11-22 05:32:26 143 ERR 20011 MOVECODE 2007-11-22 05:32:26 MOVECODE Error 20011

2007-11-22 05:32:26 143 ERR 20011 MOVECODE Bad parameter: limit source 8181 destination 8181, source [0X2AA43FE000]7863, destination [0X2AA43FE000]7873, -19

2007-11-22 05:32:26 143 ERR 20011 MOVECODE 28 bytes

2007-11-22 05:32:26 143 ERR 20011 MOVECODE + SAPDB_PascalMoveKernel.cpp:96

2007-11-22 05:32:26 143 ERR 20011 MOVECODE + 2007-11-22 05:32:26 MOVECODE Error 20013

2007-11-22 05:32:26 143 ERR 20011 MOVECODE - Module VBD600 call index 1

2007-11-22 13:33:08 40 ERR 53000 B*TREE 0200000005AF3CA0B78E010007000001

2007-11-22 13:33:08 40 ERR 53000 B*TREE Root 939069

2007-11-22 13:33:08 40 ERR 53367 B*TREE bd400_DeleteSubTrees: 865464

2007-11-26 05:33:07 165 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

2007-11-26 05:34:54 165 ERR 20011 MOVECODE vbd600.cpp:2

2007-11-26 05:34:54 165 ERR 20011 MOVECODE 2007-11-26 05:34:54 MOVECODE Error 20011

2007-11-26 05:34:54 165 ERR 20011 MOVECODE Bad parameter: limit source 8181 destination 8181, source [0X2A9BADC000]4575, destination [0X2A9BADC000]4551, -48

2007-11-26 05:34:54 165 ERR 20011 MOVECODE bytes

2007-11-26 05:34:54 165 ERR 53334 B*TREE Data page corrupted: 453698

2007-11-26 05:34:54 165 ERR 53000 B*TREE 07010000000000018EB8000000000000

2007-11-26 05:34:54 165 ERR 53000 B*TREE Index Root 379544

2007-11-26 05:34:54 165 ERR 53250 INDEX Bad Index 379544 (Root)

2007-11-26 05:34:54 165 ERR 53250 INDEX Reason "System error: BD Data page corr"

2007-11-26 05:34:54 165 ERR 20011 MOVECODE vbd600.cpp:2

2007-11-26 05:34:54 165 ERR 20011 MOVECODE 2007-11-26 05:34:54 MOVECODE Error 20011

2007-11-26 05:34:54 165 ERR 20011 MOVECODE Bad parameter: limit source 8181 destination 8181, source [0X2A9BADC000]4575, destination [0X2A9BADC000]4551, -48

2007-11-26 05:34:54 165 ERR 20011 MOVECODE bytes

2007-11-26 05:34:54 165 ERR 20011 MOVECODE + SAPDB_PascalMoveKernel.cpp:96

2007-11-26 05:34:54 165 ERR 20011 MOVECODE + 2007-11-26 05:34:54 MOVECODE Error 20013

2007-11-26 05:34:54 165 ERR 20011 MOVECODE - Module vbd600.cpp call index 2

2007-11-26 08:39:11 138 ERR 53000 B*TREE 0200000005B06A0EB88E010007000001

2007-11-26 08:39:11 138 ERR 53000 B*TREE Root 379544

2007-11-26 08:39:11 138 ERR 53367 B*TREE bd400_DeleteSubTrees: 453698

2007-11-27 05:31:43 165 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

[...]

2007-11-27 05:32:21 165 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

2007-11-27 05:32:21 165 ERR 20011 MOVECODE vbd620.cpp:1

2007-11-27 05:32:21 165 ERR 20011 MOVECODE 2007-11-27 05:32:21 MOVECODE Error 20011

2007-11-27 05:32:21 165 ERR 20011 MOVECODE Bad parameter: limit source 8192 destination 8192, source [0X2AA4502000]7041, destination [0X2AA4502000]5011, -14

2007-11-27 05:32:21 165 ERR 20011 MOVECODE 336 bytes

2007-11-27 05:32:21 165 ERR 53250 INDEX Bad Index 247269 (Root)

2007-11-27 05:32:21 165 ERR 53250 INDEX Reason "System error: Move error"

2007-11-27 05:32:21 165 ERR 51080 SYSERROR -9049 BD Primkey from inv in primtree no

2007-11-27 05:32:21 165 ERR 20011 MOVECODE vbd620.cpp:1

2007-11-27 05:32:21 165 ERR 20011 MOVECODE 2007-11-27 05:32:21 MOVECODE Error 20011

2007-11-27 05:32:21 165 ERR 20011 MOVECODE Bad parameter: limit source 8192 destination 8192, source [0X2AA4502000]7041, destination [0X2AA4502000]5011, -14

2007-11-27 05:32:21 165 ERR 20011 MOVECODE 336 bytes

2007-11-27 05:32:21 165 ERR 20011 MOVECODE + SAPDB_PascalMoveKernel.cpp:96

2007-11-27 05:32:21 165 ERR 20011 MOVECODE + 2007-11-27 05:32:21 MOVECODE Error 20013

2007-11-27 05:32:21 165 ERR 20011 MOVECODE - Module vbd620.cpp call index 1

2007-11-28 02:13:48 178 ERR 53019 CHECK Base error: index_not_accessib

2007-11-28 02:13:48 178 ERR 53019 CHECK Root pageNo: 247269

2007-11-28 02:28:27 178 ERR 53000 CHECK Check data finished unsuccessfully

2007-11-28 09:44:06 40 ERR 53000 B*TREE 0200000005B0E9D8B98E010007000001

2007-11-28 09:44:06 40 ERR 53000 B*TREE Root 247269

2007-11-28 09:44:06 40 ERR 53367 B*TREE bd400_DeleteSubTrees: 261272

[Former Member]

Hello Arne,

if you haven't already done so, I recommend that you try to find all bad indexes in DBMgui, and then drop/recreate them using DBMgui.

Regards,

Ashwath

[Former Member]

Hello Ashwath,

I already did - but the same index was marked BAD the next day.

I've already searched through the Wiki and FAQs but have not found any answer why this is happening. The database did not crash or shutdown. Can indexes be marked bad during "normal operation"?

Thanks,

Arne

[lbreddemann]

Hi Arne,

please perform a CHECK DATA of your database after you rebuild your indexes.

Usually indexes aren't marked as 'bad' in normal operation.

Another thing is - and it's just a general recommendation - update your database software. 7.6.00 is fairly old.

Go for the current 7.6.03 build if possible, as many known bugs had been fixed there.

KR Lars

[Former Member]

Hi all,

it just happened again - the same index went "BAD" after a simple SELECT:

Auto Commit: On, SQL Mode: Internal, Isolation Level: Committed
 General error;-9041 POS(1) System error: BD Index not accessible
select count(ctm_num) from cont_move where mes_num in (
  select mes_num 
  from message 
  where mty_num in (105, 109)
  and  year(mes_createdon) = 2006
  and month(mes_createdon) = 12
)

I also found some ".cor" files in the "Diagnostic Files" section of the DBMGUI. Would these point to the problem?

Every night a "util_execute VERIFY" is done in the backup script and no error is found. Can a CHECK DATA give different results?

Thanks,

Arne

[lbreddemann]

Hi Arne,

the .cor-files would help the developers to analyze the issue further. These files are just "dumps" of the bad pages.

Please try the following:

instead of recovering the indexes, drop them (all of them) and recreate them again.

Does the error reoccur?

KR Lars