On Wed, May 19, 2004 at 11:36:38AM -0400, Brent Baisley wrote:
> I wouldn't upgrade until you know where the bottleneck is (CPU, disk, 
> network, or RAM). Since you are using "professional" software, I 
> wouldn't try to change the queries. Have you made changes to your 
> my.cnf file? Since you have enough ram to hold all the data, ram is 
> probably not your bottleneck. The question is whether you have your 
> system configured to use it.

Here is the uncommented parts in the /etc/my.cnf:
[client]
port            = 3306
socket          = /tmp/mysql.sock

[mysqld]
port            = 3306
socket          = /tmp/mysql.sock
key_buffer = 384M
skip-locking
set-variable    = max_connections=3500
max_allowed_packet = 1M
table_cache = 512
sort_buffer_size = 2M
read_buffer_size = 2M
myisam_sort_buffer_size = 64M
thread_cache = 8
query_cache_size = 32M
thread_concurrency = 2
server-id       = 1

[mysqldump]
quick
max_allowed_packet = 16M

[mysql]
no-auto-rehash

[isamchk]
key_buffer = 256M
sort_buffer_size = 256M
read_buffer = 2M
write_buffer = 2M

[myisamchk]
key_buffer = 256M
sort_buffer_size = 256M
read_buffer = 2M
write_buffer = 2M

> I forget if phpbb and phpnuke use PEAR::DB to access the underlying 
> database. If they do, there is a certain amount of overhead involved 
> that you just can't get around without changing the database access 
> code to use "native" calls. Are your database and web server on the 
> same machine (a security no-no, but performance yes-yes)? If they are 
> on separate machines, make sure you have a fast connection between 
> them, at least 100MB and make sure it's full duplex.
> 
I think they don't use it, The database on a separate server, with a 1000MB LAN 
connection, Same for the 2 webservers.
I verified this now.

> 
> On May 19, 2004, at 10:34 AM, Mohammed Sameer wrote:
> 
> >On Wed, May 19, 2004 at 08:47:28AM -0400, Brent Baisley wrote:
> >>Something that small shouldn't really need optimizing. What is the 
> >>size
> >>of your data (mb?, gb?) and what does your query look like? If you are
> >>doing a wild card search on a large text field without a full text
> >>index, then those times may be the best you're going to get. Many 
> >>times
> >>it's about optimizing your query statement or you indexes rather than
> >>MySQL settings.
> >mainly we are using 2 databases:
> >675M for phpbb and 77M for phpnuke. That's not much.
> >Ok, It seems that the actual queries performed by phpbb are optimized.
> >
> >>For instance, I have a test machine (Mac G4 400Mhz, 512MB, MySQL
> >>4.0.18) that does a three table join with each table having just over
> >>100,000 rows. My initial query took about 12 seconds, but changing my
> >>query statement around a little I got it to just under 1 second. Which
> >>is pretty good considering the hardware it's on.
> >>What does you explain for the query look like? It's probably doing a
> >>full table scan, which means you're bottleneck is the disk.
> >>
> >
> >but the server is running on a duel P III with 2 GB RAM and a RAID 
> >controller.
> >Do you think there is much to be done, or shall I look for better 
> >hardware/or clustering ?
> >
> >>On May 19, 2004, at 3:26 AM, Mohammed Sameer wrote:
> >>
> >>>Hi,
> >>>I'm running a MySQL server on a duel P III 1G, with 2 GB RAM.
> >>>MySQL 4.0.18 compiled from source.
> >>>
> >>>We have 2 webservers running apache, And this is the backend database
> >>>server.
> >>>The server is really slow.
> >>>a select on a table with 138,247 rows takes about 1.6 - 2+ seconds, I
> >>>have about 50% CPU idle.
> >>>
> >>>using MyISAM table types.
> >>>
> >>>sure I can provide any statistics about the running server.
> >>>Do you think that switching to innodb'd help ? Is it a problem with
> >>>Debian woody ??
> >>>Any suggestions are welcomed.
> >>>
> >>>Many thanks!
> >>>
> >>>-- 
> >>>----------------
> >>>-- Katoob Main Developer
> >>>Linux registered user #224950, ICQ #58475622
> >>>--
> >>>Don't send me any attachment in Micro$oft (.DOC, .PPT) format please
> >>>Read http://www.fsf.org/philosophy/no-word-attachments.html
> >>>Preferable attachments: .PDF, .HTML, .TXT
> >>>Thanx for adding this text to Your signature
> >>>--
> >>>-----BEGIN GEEK CODE BLOCK-----
> >>>Version: 3.1
> >>>GCM/IT d-(++)@ s+(++):->+++ a-- C+++$>++++ UL+++$>++++ P+++$>+++++
> >>>L+++(++++)$>+++++ E>+++ W++?>$ N+>+++ o? K-? !w++ !O !M !V !PS@ !PE@
> >>>Y+ PGP=+++
> >>>t? 5? !X R? tv-- b+@ DI D+ G-- e++>+++ h-->++ !r y?
> >>>------END GEEK CODE BLOCK------
> >>>
> >>-- 
> >>Brent Baisley
> >>Systems Architect
> >>Landover Associates, Inc.
> >>Search & Advisory Services for Advanced Technology Environments
> >>p: 212.759.6400/800.759.0577
> >>
> >
> >-- 
> >----------------
> >-- Katoob Main Developer
> >Linux registered user #224950, ICQ #58475622
> >--
> >Don't send me any attachment in Micro$oft (.DOC, .PPT) format please
> >Read http://www.fsf.org/philosophy/no-word-attachments.html
> >Preferable attachments: .PDF, .HTML, .TXT
> >Thanx for adding this text to Your signature
> >--
> >-----BEGIN GEEK CODE BLOCK-----
> >Version: 3.1
> >GCM/IT d-(++)@ s+(++):->+++ a-- C+++$>++++ UL+++$>++++ P+++$>+++++
> >L+++(++++)$>+++++ E>+++ W++?>$ N+>+++ o? K-? !w++ !O !M !V !PS@ !PE@ 
> >Y+ PGP=+++
> >t? 5? !X R? tv-- b+@ DI D+ G-- e++>+++ h-->++ !r y?
> >------END GEEK CODE BLOCK------
> >
> -- 
> Brent Baisley
> Systems Architect
> Landover Associates, Inc.
> Search & Advisory Services for Advanced Technology Environments
> p: 212.759.6400/800.759.0577
> 

-- 
----------------
-- Katoob Main Developer
Linux registered user #224950, ICQ #58475622
--
Don't send me any attachment in Micro$oft (.DOC, .PPT) format please
Read http://www.fsf.org/philosophy/no-word-attachments.html
Preferable attachments: .PDF, .HTML, .TXT
Thanx for adding this text to Your signature
--
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GCM/IT d-(++)@ s+(++):->+++ a-- C+++$>++++ UL+++$>++++ P+++$>+++++
L+++(++++)$>+++++ E>+++ W++?>$ N+>+++ o? K-? !w++ !O !M !V !PS@ !PE@ Y+ PGP=+++
t? 5? !X R? tv-- b+@ DI D+ G-- e++>+++ h-->++ !r y?
------END GEEK CODE BLOCK------

Attachment: signature.asc
Description: Digital signature

Reply via email to