Hi,
We're chaning it to INT(9). Apparently someone remembered to change the type
of data in this field from an alphanumeric value to an INT(9).
I'm going to change this asap.
Thanks
BR
AJ
On Mon, Sep 6, 2010 at 5:17 AM, mos mo...@fastmail.fm wrote:
At 04:44 AM 9/3/2010, Alexandre Vieira
Hi List,
In a 20m interval in our max load I have:
OS WAIT ARRAY INFO: reservation count 637, signal count 625
Mutex spin waits 0, rounds 19457, OS waits 428
RW-shared spins 238, OS waits 119; RW-excl spins 13, OS waits 8
(The values are the difference between the start and end of this 20m
On 9/3/2010 3:15 PM, Johnny Withers wrote:
It seems that when your index is PRIMARY on InnoDB tables, it's magic and is
part of the data thereby it is not included in the index_length field.
I have never noticed this. I don't think adding a new index will make a
difference.
You could try
At 04:44 AM 9/3/2010, Alexandre Vieira wrote:
Hi Johnny,
mysql EXPLAIN SELECT * FROM clientinfo WHERE userid='911930694';
++-++---+---+-+-+---+--+---+
| id | select_type | table | type | possible_keys | key |
Hi Johnny,
mysql EXPLAIN SELECT * FROM clientinfo WHERE userid='911930694';
++-++---+---+-+-+---+--+---+
| id | select_type | table | type | possible_keys | key | key_len
| ref | rows | Extra |
On 02/09/2010 6:05 p, Alexandre Vieira wrote:
Hi Jangita,
I'm 15779 innodb_buffer_pool_pages_free from a total of 22400. That's
246MB of 350MB free.
| Innodb_buffer_pool_pages_data | 6020 |
| Innodb_buffer_pool_pages_dirty| 1837 |
| Innodb_buffer_pool_pages_flushed |
Ok, so I'm stumped?
What kind of hardware is behind this thing?
-JW
On Fri, Sep 3, 2010 at 4:44 AM, Alexandre Vieira nul...@gmail.com wrote:
Hi Johnny,
mysql EXPLAIN SELECT * FROM clientinfo WHERE userid='911930694';
Hi,
I've done some tests with INT(8) vs the VARCHAR(23) on the userid PK and it
makes a little difference but not enough for the application to run in real
time processing.
It's a Sun Fire V240 2x 1.5ghz UltraSparc IIIi with 2GB of RAM.
MySQL is eating 179MB of RAM and 5,4% of CPU.
PID
What does
SHOW TABLE STATUS LIKE 'table_name'
Say about this table?
-JW
On Fri, Sep 3, 2010 at 8:59 AM, Alexandre Vieira nul...@gmail.com wrote:
Hi,
I've done some tests with INT(8) vs the VARCHAR(23) on the userid PK and it
makes a little difference but not enough for the application to
Hi,
mysql SHOW TABLE STATUS LIKE 'clientinfo';
Very confusing...
Why is index_length zero ?
On top of that, there's only 500K rows in the table with a data size of
41MB. Maybe InnoDB is flushing to disk too often?
What's the output of iostat -dxk 60 ? (run for a minute+ to get 2 output
girds)
--
*Johnny
Hi,
The DB is working on /var, which is md2 / md12 / md22.
extended device statistics
device r/sw/s kr/s kw/s wait actv svc_t %w %b
md2 0.1 80.00.4 471.4 0.0 1.0 12.2 0 94
md10 0.05.70.0 78.8 0.0 0.1 19.7 0 9
I think your MySQL instance is disk bound.
If you look at your iostats, md2, 12 and 22 have a ~10ms wait time before a
request can be processed. iostat is also reporting those disks are 75%+
utilized which means they are doing about all they can do.
Anyway you can add more disks? Add faster
Hi,
When creating a table in MySQL with a PK it automatically creates an INDEX,
correct?
The Index_Length: 0 is rather strange..I've created a new INDEX on top of my
PK column on my test system and Index_Length shows a big value different
from 0. Do you think this might have any impact?
mysql
It seems that when your index is PRIMARY on InnoDB tables, it's magic and is
part of the data thereby it is not included in the index_length field.
I have never noticed this. I don't think adding a new index will make a
difference.
You could try moving your log files to a different disk array
Hi list,
I'm having some performance problems on my 5.0.45-log DB running on Solaris
8 (V240).
We only have one table and two apps selecting, updating, inserting and
deleting massively and randomly from this table.
The table is very simple. All SELECTs,INSERTs,UPDATEs and DELETEs have only
one
Can you show us the table structure and sample queries?
On Thursday, September 2, 2010, Alexandre Vieira nul...@gmail.com wrote:
Hi list,
I'm having some performance problems on my 5.0.45-log DB running on Solaris
8 (V240).
We only have one table and two apps selecting, updating, inserting
What is the hardware spec? Anything else running on the box?
Why are you replicating but not making use of the slave?
Can you post the output of SHOW CREATE TABLE?
Regards
John
On 2 September 2010 12:50, Alexandre Vieira nul...@gmail.com wrote:
Hi list,
I'm having some performance problems
John, Johnny,
Thanks for the prompt answer.
mysql SHOW CREATE TABLE clientinfo;
On 02/09/2010 4:46 p, Alexandre Vieira wrote:
John, Johnny,
Thanks for the prompt answer.
...
We also run some other applications in the server, but nothing that consumes
all the CPU/Memory. The machine has almost 1GB of free memory and 50% of
idle CPU time at any time.
TIA
BR
Alex
Hi Jangita,
I'm 15779 innodb_buffer_pool_pages_free from a total of 22400. That's 246MB
of 350MB free.
| Innodb_buffer_pool_pages_data | 6020 |
| Innodb_buffer_pool_pages_dirty| 1837 |
| Innodb_buffer_pool_pages_flushed | 673837 |
| Innodb_buffer_pool_pages_free |
Cc: mysql@lists.mysql.com
Subject: Performance problems on MySQL
John, Johnny,
Thanks for the prompt answer.
mysql SHOW CREATE TABLE clientinfo
for your
delete queries?
DELETE FROM clientinfo WHERE units='155618918';
-Original Message-
From: Alexandre Vieira [mailto:nul...@gmail.com]
Sent: Thursday, September 02, 2010 8:46 AM
To: John Daisley; joh...@pixelated.net
Cc: mysql@lists.mysql.com
Subject: Performance problems on MySQL
Hello All,
I've recently installed mySQL on a RH7.2 box without any problems - it runs
great. I've been testing some long running queries (full table scans, etc)
under different scenarios, and get wildy differing results. Between each test
case, I was flushing all tables and re-starting
On Tue, Mar 12, 2002 at 11:37:20AM -0800, Henry Hank wrote:
Hello All,
I've recently installed mySQL on a RH7.2 box without any problems
- it runs great. I've been testing some long running queries (full
table scans, etc) under different scenarios, and get wildy differing
results.
25 matches
Mail list logo
