Hi all,

I'm using load testing tools to simulate concurrent access to a 4-datanode MySQL
Cluster.

Following are two statement in my concurrent access: 1> select * from test_table where blogid=xxx and valid='0' order by lastbuilddate limit 7;

2> update test_table set pvnum=pvnum+1 where id=123;

table structure:

create table test_table{ id biginit(20) not null auto_increment, pvnum int(8) default '0', blog_id int(8) not null default '0', valid int(4) default null, lastbuilddate datetime, primary key (id) }

When I used a large amount concurrent access , there are one or two statement 1>
can not finish. I wonder if it's a lock problem. Because too many "select" are waiting for
statement 2> to release the row lock in row where id=123 . Could some queries overflow the waiting queue ?

And if so, how can I resolve the problem? Can I increase waiting queue space?

Best Regards

------------------- Technical Support & Service Turbolinux Inc.,China Tel:+86-010 6505 4020 Fax:(86-010) 65054017