Description:
Hi,
I've been reading up on a similar issue and I'm not sure whether it's a simple permissions problem. I'm running mysql v 5.4.1-beta and 5.4.0-beta over NFS.
I obtain the following error even though the folders are owned by the same user under which the db instance is running. Please see the log below. The database instance keeps restarting after each attempted table update.
There have been bug reports regardning fsync with both the InnoDB storage engine and NFS itself. The mysql/innodb bugs and patches have apparently been applied to versions 5.1x and 6.x. However, they have not seen the light of day in the latest 5.4x branches.
== related bug reports ==
http://bugs.mysql.com/bug.php?id=46718 (mysql on ppc)
http://bugs.mysql.com/bug.php?id=34823 (mysql, innodb)
http://www.freebsd.org/cgi/query-pr.cgi?pr=86944 (freebsd related NFS issue)
== system information ==
SunOS 5.10 Generic_139555-08 sun4v sparc SUNW,SPARC-Enterprise-T5120 Solaris
NFS4 client
== log below ==
090908 13:47:03 InnoDB: Log file ./ib_logfile0 did not exist: new to be created
InnoDB: Setting log file ./ib_logfile0 size to 256 MB
InnoDB: Database physically writes the file full: wait...
InnoDB: Progress in MB: 100 200
090908 13:47:32 InnoDB: Error: the OS said file flush did not succeed
090908 13:47:32 InnoDB: Operating system error number 13 in a file operation.
InnoDB: The error means mysqld does not have the access rights to
InnoDB: the directory.
InnoDB: File operation call: 'flush'.
InnoDB: Cannot continue operation.
How to repeat:
unpack the tarballs onto an NFS-served location
cd into the base folder
issue scripts/mysql_install_db
Suggested fix:
See the body of my message for patches that haven't been applied to the 5.4.x-beta brances
