MySQL Lists are EOL. Please join:

List:General Discussion« Previous MessageNext Message »
From:Ying Zhang Date:June 23 1999 3:30pm
Subject:Re: Benchmarking my MySQL db ...
View as plain text  
It might be how the kernel was compiled on Redhat 6.0.  If you're using IDE
drives, try running this:

# /sbin/hdparm -Tt /dev/hda

That'll give you basic hard drive benchmarks.  Next, run

# /sbin/hdparm -c1 -d1 -u1 /dev/hda

That enables UDMA and 32bit access and unmask IRQ

Run the benchmark again

# /sbin/hdparm -Tt /dev/hda

And you should see dramatic improvements.  If that works out, run /sbin/hdparm
-c1 -d1 -u1 /dev/hda from your init scripts (/etc/rc.d/rc.local is a good
place).  Your disk IO performance should increase dramatically.

Ying


Tom Cunningham wrote:

> I'm benchmarking a new machine to serve as an apache / mysql server. After
> running the MySQL benchmarks (connect, wisconsin, etc), I've determined
> that the new machine is twice as fast as the existing machine. However,
> when testing a database-driven CGI on the two machines, the old machine
> outperforms the new machine...
>
> >From Apache Benchmark :
>                         Old Machine             New Machine
> Requests per second :   1.47                    .84
> Processing Min          1980                    1940
> Processing Avg          2665                    4685
> Processing Max          4273                    6869
>
> I ftp'ed the .ISM, .ISD, and .frm files from the old machine onto the new
> machine, so it can't be differences in the tables on the two machines. The
> old machine is running Red Hat 5.2, the new machine Red Hat 6.0.
> Mysql versions differ on the two machines. The old machine runs
> mysql 3.22.14-gamma, while the new one runs Mysql 3.22.23b. Versions
> of Perl, DBI, Apache, and mod-perl are all identical. I have noticed the
> following though..
>
> [root@newmachine /root]# /usr/local/bin/mysqladmin -u root -pnew processlist
> +------+------+-----------+----+-----------+------+-------+------+
> | Id   | User | Host      | db | Command   | Time | State | Info |
> +------+------+-----------+----+-----------+------+-------+------+
> | 1222 | root | localhost |    | Processes | 0    |       |      |
> +------+------+-----------+----+-----------+------+-------+------+
>
> [root@oldmachine ]# /usr/local/bin/mysqladmin -u root -pold processlist
> +-----+--------+-----------+------------+-----------+------+-------+------+
> | Id  | User   | Host      | db         | Command   | Time | State | Info
> |
> +-----+--------+-----------+------------+-----------+------+-------+------+
> | 108 | user_x | localhost | production | Sleep     | 452  |       |
> |
> | 109 | user_x | localhost | production | Sleep     | 456  |       |
> |
> | 110 | user_x | localhost | production | Sleep     | 454  |       |
> |
> | 111 | user_x | localhost | production | Sleep     | 507  |       |
> |
> | 112 | user_x | localhost | production | Sleep     | 453  |       |
> |
> | 113 | user_x | localhost | production | Sleep     | 494  |       |
> |
> | 116 | user_x | localhost | production | Sleep     | 452  |       |
> |
> | 117 | user_x | localhost | production | Sleep     | 453  |       |
> |
> | 119 | user_x | localhost | production | Sleep     | 488  |       |
> |
> | 120 | user_x | localhost | production | Sleep     | 461  |       |
> |
> | 123 | root   | localhost |            | Processes | 0    |       |
> |
> +-----+--------+-----------+------------+-----------+------+-------+------+
>
> AND
>
> [root@newmachine /root]# ps ax | grep mysql
> 12205 ?        S      0:00 sh /usr/local/bin/safe_mysqld -O
> key_buffer=100M
> -O r
> 12216 ?        SN     0:00 /usr/local/libexec/mysqld --basedir=/usr/local
> --data
> 12218 ?        SN     0:00 /usr/local/libexec/mysqld --basedir=/usr/local
> --data
> 12219 ?        SN     0:00 /usr/local/libexec/mysqld --basedir=/usr/local
> --data
> 16504 pts/1    S      0:00 grep mysql
>
> [root@oldmachine] ~ 10 >> ps ax | grep mysql
> 13726  p0 S    0:00 sh /usr/local/bin/safe_mysqld -O key_buffer=100M -O
> record_b
> 13737  p0 S N  0:00 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13739  p0 S N  0:00 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13740  p0 S N  0:00 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13973  p0 S N  0:07 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13974  p0 S N  0:17 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13975  p0 S N  0:16 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13976  p0 S N  0:03 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13980  p0 S N  0:17 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13981  p0 S N  0:04 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13984  p0 S N  0:08 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13985  p0 S N  0:09 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13987  p0 S N  0:02 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13988  p0 S N  0:09 /usr/local/libexec/mysqld --basedir=/usr/local/
> --datadir=/u
> 13999  p4 S    0:00 grep mysql
>
> Anyone have any suggestions as to what might be the problem? I really find
> it strange that the new machine can best the old machine in the
> "Processing Minimum time" category, and still lose badly because of high
> "Processing Maximum times". What could cause such wide varieties in query
> speed? Are there setup options I can tune that might fix this?
>
> Thanks,
>
> Tom
>
> ---------------------------------------------------------------------
> Please check "http://www.mysql.com/Manual_chapter/manual_toc.html" before
> posting. To request this thread, e-mail mysql-thread5674@stripped
>
> To unsubscribe, send a message to the address shown in the
> List-Unsubscribe header of this message. If you cannot see it,
> e-mail mysql-unsubscribe@stripped instead.

Thread
Benchmarking my MySQL db ...Tom Cunningham23 Jun
  • Re: Benchmarking my MySQL db ...Ying Zhang23 Jun