RE: MySQL 5.0.67 on SMP
Yes，I tested from two mysql clients with your code and it showed exactly as
you pointed out.
The top from the mysqld machine showed:
Cpu0 : 100.0% us, 0.0% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.0%
si
Cpu1 : 0.3% us, 0.0% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu2 : 0.0% us, 0.3% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu3 : 100.0% us, 0.0% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.0%
si
Mem: 2074824k total, 977252k used, 1097572k free,74844k buffers
Swap: 2031608k total,0k used, 2031608k free, 720236k cached
24022 mysql 25 0 100 1:59.25 2.7 554m 55m 2448 R mysqld
24027 mysql 25 0 100 2:12.25 2.7 554m 55m 2448 R mysqld
Thank you again.
Yours,
Xu Feng
> -Original Message-
> From: Dan Nelson [mailto:dnel...@allantgroup.com]
> Sent: 2008年12月16日 13:51
> To: xufeng
> Cc: mysql@lists.mysql.com
> Subject: Re: MySQL 5.0.67 on SMP
>
> In the last episode (Dec 16), xufeng said:
> > Is there a way to check if my MySQL5.0.67 works well on SMP?
> > I have two CPUs with each two cores, and I want to know if MySQL
distributes
> > loads over the two CPUs.
> > System OS: Linux 2.6.9-42.ELsmp
> > MySQL Version: 5.0.67
> > Intel(R) Xeon(TM) CPU 3.00GHz * 2
>
> The easiest way to check would be to run top, then run a long-running
> CPU-hungry query like
>
> SELECT BENCHMARK(1,ENCODE('hello','goodbye'));
>
> from two separate mysql client sessions. You should see two CPUs worth
> of load on the system at that point.
>
> In fact, any version of mysql should scale to multiple CPUs as long as
> your OS supports kernel-based threads (most do). Note that a single
> query will always only use one CPU, so you need multiple queries in
> parallel to use more.
>
> --
> Dan Nelson
> dnel...@allantgroup.com
>
> --
> MySQL General Mailing List
> For list archives: http://lists.mysql.com/mysql
> To unsubscribe:http://lists.mysql.com/mysql?unsub=xuf...@yuanjie.net
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/mysql?unsub=arch...@jab.org
full text search ft_min_word_len
Hi, When I use ft_min_word_len=3 under [mysqld] section, mysql fails to start. ft_min_word_len=3 parameter has been added under [myisamchk] only , which works. mysql Ver 14.12 Distrib 5.0.67, for redhat-linux-gnu (i686) using readline 5.1 Anyone know about this?
Re: MySQL 5.0.67 on SMP
Hi all, xufeng wrote: > Hi fire9, > Thank you for your reply. > Do you have some official document on "MySQL 5.0.67 not supporting SMP"? > If so, would you please give me some info on how to get the smp patches? > > Yours, > Xu Feng > > > > > From: fire9 [mailto:fire9di...@gmail.com] > Sent: 2008?12?16? 11:54 > To: xufeng > Subject: Re: MySQL 5.0.67 on SMP > > hi, > I think the version is not support SMP.You may use smp patches in this > version. > ? 2008-12-16,??11:44, xufeng ??: > > > Hi > > Is there a way to check if my MySQL5.0.67 works well on SMP? > I have two CPUs with each two cores, and I want to know if MySQL distributes > loads over the two CPUs. > System OS: Linux 2.6.9-42.ELsmp > MySQL Version: 5.0.67 > Intel(R) Xeon(TM) CPU 3.00GHz * 2 that idea of "MySQL 5.0.67 not supporting SMP" is plain wrong: The MySQL server uses threads internally, one thread per connection, and any decent operating system (Linux is among them) will distribute the threads across the available CPUs and cores. There are just two factors limiting this: a) "One thread per connection" means that you need multiple connections with active commands to profit from your SMP system. As long as only one applications sends a single stream of commands to the MySQL server, it cannot use the multiple CPUs / cores. b) If several of these threads need to access the same resource (typically, some data structure), they will synchronize, and only one thread will proceed immediately while the other(s) block(s). If that synchronization covers a very large block of code, the performance of the system will not scale proportionally to the number of CPUs / cores. Regards, Jörg -- Joerg Bruehe, MySQL Build Team, joerg.bru...@sun.com Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten Geschaeftsfuehrer: Thomas Schroeder, Wolfgang Engels, Dr. Roland Boemer Vorsitzender des Aufsichtsrates: Martin Haering Muenchen: HRB161028 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/mysql?unsub=arch...@jab.org
Re: MySQL 5.0.67 on SMP
Hi, I think the known problems with MySQL and SMP architecture are more based on storage engine restrictions, so InnoDB for example does not scale well beyond say 8 cores. There are a number of patches to help solve some of these scaling problems. http://www.percona.com/docs/wiki/release:start http://code.google.com/p/google-mysql-tools/ You can read about these here http://www.mysqlperformanceblog.com/2008/10/20/improved-innodb-rw_lock-patch/ Ewen On Tue, Dec 16, 2008 at 4:44 AM, xufeng wrote: > Hi > > Is there a way to check if my MySQL5.0.67 works well on SMP? > I have two CPUs with each two cores, and I want to know if MySQL distributes > loads over the two CPUs. > System OS: Linux 2.6.9-42.ELsmp > MySQL Version: 5.0.67 > Intel(R) Xeon(TM) CPU 3.00GHz * 2 > > Thank you in advance. > > Yours, > Xu Feng > > > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe:http://lists.mysql.com/mysql?unsub=ewen.fort...@gmail.com > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/mysql?unsub=arch...@jab.org
