Re: innodb_file_per_table cost

2009-05-28 Thread Sebastien Moretti
Hi Sebastian, It depends. In general, no. In some filesystems and operating systems, it actually helps. I think you can base your decision on whether it makes server administration easier for you. Regards Baron Thanks It seems there are no clear thresholds between I/O access, the number

RE: innodb_file_per_table cost

2009-05-28 Thread Jerry Schwartz
-Original Message- From: Sebastien Moretti [mailto:sebastien.more...@unil.ch] Sent: Thursday, May 28, 2009 2:38 AM To: Baron Schwartz; mysql@lists.mysql.com Subject: Re: innodb_file_per_table cost Hi Sebastian, It depends. In general, no. In some filesystems and operating systems

RE: innodb_file_per_table cost

2009-05-28 Thread Rolando Edwards
[mailto:sebastien.more...@unil.ch] Sent: Thursday, May 28, 2009 2:38 AM To: Baron Schwartz; mysql@lists.mysql.com Subject: Re: innodb_file_per_table cost Hi Sebastian, It depends. In general, no. In some filesystems and operating systems, it actually helps. I think you can base your

Re: innodb_file_per_table cost

2009-05-28 Thread Baron Schwartz
[JS] I strongly suspect that MySQL, like any other random access, variable record length scheme, would find it easier to manage the internal layout of separate files. The rows would tend more to be of similar sizes, leading to less obnoxious fragmentation, and the files themselves would be

Re: innodb_file_per_table cost

2009-05-27 Thread Baron Schwartz
Hi Sebastian, It depends. In general, no. In some filesystems and operating systems, it actually helps. I think you can base your decision on whether it makes server administration easier for you. Regards Baron On Wed, May 27, 2009 at 2:45 AM, Sebastien MORETTI sebastien.more...@unil.ch