I have a table with over 26 million records that experiences a high volume of traffic.

Recently, query results became a little screwy; upon investigation it was found that 
the table and the index were fouled up.

The results of an isamchk on the table yielded:

Checking ISAM file: /usr8/hydrology/measurement
Data records: 27034726   Deleted blocks:       0
- check file-size
./bin/isamchk: warning: Size of datafile is: 11             Should be: 3554138
- check delete-chain
- check index reference
- check data record references index: 1
./bin/isamchk: error: Found key at page 1024 that points to record outside datafile
ISAM-table '/usr8/hydrology/measurement' is corrupted
Fix it using switch "-r" or "-o"

and on the index yielded:

Data records: 27034726   Deleted blocks:       0
- check file-size
./bin/isamchk: warning: Size of datafile is: 11             Should be: 3554138
- check delete-chain
- check index reference
- check data record references index: 1
./bin/isamchk: error: Found key at page 1024 that points to record outside datafile
ISAM-table '/usr8/hydrology/measurement.ISM' is corrupted
Fix it using switch "-r" or "-o"


Trying to "fix" the table as suggested results in the datafile being reduced to a file 
size of 0 bytyes.  Before the fix it is over 4Gigs.  

I have tried several different options for isamchk including the --quick,which does 
not kill the data file but still does not work.

So the question is:

since the isamchk can detect the 27 million records, what is my best hope for getting 
them all out.

select * results in garbage

any thoughts would be appreciated.

I can get about 90% of the records out with carefully crafted where clauses, 
but I am a little concerned with the integrity of them.  What is the probability 
that the integrity of these records can be trusted.  They "look" ok and a few random 
checks that I have done show no problems.

thanks to all.

Kevin Kotun, P.E.
Hydrologist
Everglades National Park





---------------------------------------------------------------------
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/           (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Reply via email to