From: Michael Widenius
Date: June 15 2001 10:18am
Subject: Re: MySQL 3.23.39 released
List-Archive: http://lists.mysql.com/mysql/76680
Message-Id: <15145.57570.682977.709953@narttu.mysql.fi>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit


Hi!

>>>>> "Steven" == Steven Roussey <sroussey@stripped> writes:

Steven> Changes in release 3.23.39
Steven> --------------------------

Steven>    * Fixed problem with shutdown when `INSERT DELAYED' was waiting for
Steven>      a `LOCK TABLE'.

Steven>    * Running `myisamchk --fast --force' will no longer repair tables
Steven>      that only had the open count wrong.
Steven> --------------------------

Steven> THANK YOU! You don't know what it was for the first item to require a forced
Steven> shutdown that made myisamchk take a huge amount of time. Although we have
Steven> not had the need for a `myisamchk --fast --force' in a while, I tell you if
Steven> you had a big fulltext index, you would avoid it like the plague.

Steven> Now I have a question: Does this also apply to automatic repair inside
Steven> mysqld? I'd finally use the feature if its true...


If you just run CHECK TABLE, it will repair the open count (it should
always have done that).

myisamchk --fast --force is equal to:

run CHECK TABLE on all tables
Run REPAIR TABLE on tables which was not marked as ok.

Regards,
Monty