Bug #29693 | Setting key_cache_block_size=4096 corrupts key file | ||
---|---|---|---|
Submitted: | 10 Jul 2007 17:08 | Modified: | 13 Jul 2007 22:52 |
Reporter: | James Blair | Email Updates: | |
Status: | Can't repeat | Impact on me: | |
Category: | MySQL Server: MyISAM storage engine | Severity: | S2 (Serious) |
Version: | 5.0.41 | OS: | Linux |
Assigned to: | CPU Architecture: | Any | |
Tags: | corruption, key_cache_block_size |
[10 Jul 2007 17:08]
James Blair
[11 Jul 2007 11:14]
Sveta Smirnova
Thank you for the report. We can not repeat described behaviour on our side. Additionally version 5.0.38 is old. Pleas upgrade to current version 5.0.44 of 5.0.41, try with it and if you are able to repeat error provide output of SHOW TABLE STATUS LIKE 'rejects'
[11 Jul 2007 16:12]
James Blair
I installed the RPMS from: http://dev.mysql.com/downloads/mysql/5.0.html#linux-rhel5-x86-64bit-rpms on an EMT64 machine running RHEL5. The client now displays: Server version: 5.0.41-community MySQL Community Edition (GPL) I executed the procedure I described previously, and got this result: mysql> show warnings\G *************************** 1. row *************************** Level: Error Code: 126 Message: Incorrect key file for table './test/rejects.MYI'; try to repair it 1 row in set (0.00 sec) mysql> SHOW TABLE STATUS LIKE -> 'rejects'\G *************************** 1. row *************************** Name: rejects Engine: MyISAM Version: 10 Row_format: Dynamic Rows: 0 Avg_row_length: 0 Data_length: 0 Max_data_length: 281474976710655 Index_length: 11264 Data_free: 0 Auto_increment: NULL Create_time: 2007-07-11 09:09:40 Update_time: 2007-07-11 09:09:40 Check_time: NULL Collation: latin1_swedish_ci Checksum: NULL Create_options: Comment: 1 row in set (0.00 sec)
[11 Jul 2007 16:18]
James Blair
my.cnf: [mysqld] datadir=/var/lib/mysql socket=/var/lib/mysql/mysql.sock key_cache_block_size = 4096 [mysql.server] user=mysql [mysqld_safe] err-log=/var/log/mysqld.log pid-file=/var/run/mysqld/mysqld.pid
[13 Jul 2007 13:43]
Sveta Smirnova
Thank you for the feedback. I could not repeat described behaviour with indicated binaries on RHEL5. So I close the report as "Can't repeat". If you have new idea how to repeat the problem feel free to provide it and reopen the report.
[13 Jul 2007 22:52]
James Blair
Did you set key_cache_block_size=4096? I produced the problem on three versions of MySQL, on three different machines, under two different OS distributions. I'm convinced it's repeatable.
[15 Jul 2007 20:01]
MySQL Verification Team
I also cannot repeat corruption using 5.0.46BK. I suspect this bug in a duplicate of one of the following: bug #19079 (key_cache_block_size set to 4K corrupts indexes) bug #25853 (key_cache_block_size=4MB causes table corruptions + duplicate key errors)
[15 Jul 2007 21:02]
MySQL Verification Team
Just wanted to confirm that 5.0.44 still has this bug. The latest source of 5.0.46 built from BitKeeper repositories shows that it's been fixed already. So when 5.0.46 is officially released, it will have the fix.