Bonjour,
When I insert a lot of rows in a table from another table in MyISAM
format it seems that indexes in the destination table are not
updated.
For example in table A I have index 1. This table contains many
rows.
I have table B with same structure as table A but
Bonjour,
I configured two of my servers for replication.
On the master I added :
log-bin
master-host = xxx..
master-user = x
master-password = x
master-port = 3306
server-id = 14
master-connect-retry=10
binlog-do-db=db1
binlog-do-db=db2
Heikki,
HT since you are French, I assume you have accent characters in the strings you
HT have stored into your database.
HT The ordering of accent characters changed in 3.23.44!
No accents.
InnoDB tables on this server are mainly used for statistics,
counters etc ...
HT You
HT 616e647265747469 andretti
HT 616e6472e9737069 andr.spi
HT there is a character e9 in your string. It is not an ordinary character and
HT can cause an error in the sorting order between InnoDB-.41 and InnoDB-.49.
ok :)
mysql insert into popup select * from popup2;
ERROR 1213: Deadlock
Bonjour,
mysql Ver 11.15 Distrib 3.23.47, for sun-solaris2.8 (sparc)
It crashed after more than a month with no problem and here are the logs :
020225 11:04:48 mysqld started
020225 11:04:50 InnoDB: Started
/usr/local/mysql/libexec/mysqld: ready for connections
020226 12:01:05 failed
Bonjour,
I have a problem with Mysql and access rights :
I have a user who has rights to access tables in the database
'membres' :
#/ mysqlaccess 192.168.0.1 myuser membres
Could not open outputfile
Bonjour,
I have a problem with Mysql and access rights :
I have a user who has rights to access tables in the database
'membres' :
#/ mysqlaccess 192.168.0.1 myuser membres
Could not open outputfile
Bonjour,
Did you try to connect to the server from network ?
Try to force mysql client not to use the socket (what is does when
it's connecting to localhost).
Try to connect from another server or even from the same server but
with the --host=xxx.xxx.xxx.xxx where
Bonjour,
Any idea ?
I will try to make the server crash one more time and tell you if I
find the query.
Regards,
Alex.
020108 9:43:58 InnoDB: Started
/opt/mysql/libexec/mysqld: ready for connections
InnoDB: Error: undo-id is 2860816
InnoDB: Assertion failure in thread 54 in file
Bonjour,
I have a table :
[11:09am] mysqldump -d cotations historique_AMS
# MySQL dump 8.16
#
# Host: localhostDatabase: cotations
#
# Server version3.23.45
#
# Table structure for table 'historique_AMS'
#
CREATE TABLE
in set (27.02 sec)
Regards,
Alex.
==
RE BAUMEISTER Alexandre wrote:
Bonjour,
I have a table :
[11:09am] mysqldump -d cotations historique_AMS
# MySQL dump 8.16
#
# Host: localhostDatabase: cotations
Bonjour Sinisa,
SM There were some bug fixes with packed keys.
SM 3.23.47 will soon be out and it would be fine if you could test your
SM case with it again.
I tried to do
alter table historique_AMS PACK_KEYS=0;
But it did not correct the problem.
Then I wanted check if the
Bonjour,
BA But I forgot the -d ... and here is what I found : There are some
BA \r at the end of some symbols !
I corrected everything with one query :
mysql UPDATE historique_AMS set symbole=TRIM(TRAILING '\r' FROM symbole);
These '\r' may come from a program a colleague did.
Bonjour,
I've rebooted my server with sync;sync;reboot. It seems that
Mysql/InnoDB did not have enough time for shutting down properly
because when the server restarted I had these messages in the error
log of Mysql :
InnoDB: Doing recovery: scanned up to log sequence number
Bonjour,
I have a table created with :
CREATE TABLE mybuffer_NAS_D (
id int(10) unsigned NOT NULL auto_increment,
service varchar(10) NOT NULL default '',
date datetime NOT NULL default '-00-00 00:00:00',
data text,
PRIMARY KEY (id),
KEY id_date (date)
) TYPE=InnoDB;
Bonjour,
With 3.22 I used to do some queries like DELETE FROM TABLE. And
this was very rapid as Mysql only deletes and re-create the files.
I knew that DELETE was very long with InnoDB because it has to
clear all the table space used by the table from which you DELETE.
But
Bonjour,
I have a problem with Mysql-3.23.39 and users access rights. It's
the second time I notice this problem since I installed .39 on two
of my servers. But yesterday, the bug was even more flagrant.
I had a user which had all rights on a database when connecting from
Sinisa,
mysql show grants for news@'192.168.1.%';
++
| Grants for [EMAIL PROTECTED]%
|
Sinisa,
| GRANT USAGE ON *.* TO 'news'@'192.168.1.%' IDENTIFIED BY PASSWORD
'' |
| GRANT SELECT ON cotations.* TO 'news'@'192.168.1.%'
|
| GRANT SELECT, INSERT, UPDATE, DELETE ON contenu.* TO 'news'@'192.168.1.%'
|
|
Stanislav,
HT I have a question about using InnoDB format of tables with MySQL.
HT We have a following problem:
HT our application collects data into MySQL database during some period of time
HT (say 1 hour) and then we need to analyze these data then move insert results
HT of analysis into
SM I am sending a patch straight to you
Do someone see improvements compiling Mysql with Gcc 3.0 ?
Regards,
Alex.
-
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
Bonjour Werner,
My web site as a lot of hits (really!), it's load balanced on the
web side on more than 30 servers.
I've to save database load as much as possible.
The problem with your implementation is that it needs at least one
more insert query for each page or even more if
Bonjour Werner,
Am I naive or is this a viable proposition?
WS Oh, I see, you would like to connect this information with the
WS pid, right? I wouldn't know how to do that. How do you get a pid?
WS Is this something MySQL could provide?
No, not with the pid.
With the number of places
The only thing I would like is to be able to attach a comment to a
connection id which could be returned by the mysqladmin process
command. And something that would not add more load to Mysql nor
more coding in the client.
AB Yes!! Great idea. This would be helpful for
Bonjour Justin,
user agent was just an example of what the commentary could be.
You could put what you want as long as it's a string.
I agree with you and I already had this reflection that it could be
a modification that could create some problems to some.
Instead of modifying
Bonjour Justin,
JF I don't know why I got interested in this thread...
More important (or less), why are you so negative ?
JF I suppose you don't need a change to the wire protocol if
JF you have, logically, something like below. The API has
JF to be changed, however.
JF id =
Bonjour Scott,
SH Beyond that, you can already accomplish 90% of what you'd want by
SH simply coding your queries as /* Comment */ SELECT ..., and so
SH on. For connections which are not currently running a query, this
SH won't work - but in my experience, you usually don't care much
SH
Bonjour,
When you have a big web site with a lot of pages, programs,
cronjobs, applications ... connecting to Mysql, it's sometimes
difficult to find from where a given query/connection comes from.
And it's sometimes urgent to find the origin of a disturbing query.
Bonjour Heikki,
In fact this morning Mysql is ... crashed on this server :(
It never crashed with 3.23.38 :(
I don't yet know if this is really a 3.23.39 problem or something
else.
If I'm able to restart Mysql, I will downgrade to 3.23.38 in order
to see if it's back to
Bonjour,
Solaris 2.8 on SPARC 4x64bit processor .
Mysql-3.23.37 with InnoDB
What I was doing : I was altering a lot of table, converting them
from InnoDB to MyISAM type (I don't know if this is the cause of my
problem ...)
Now, on the console here is what happens :
[4:52pm]#
Bonjour,
Sorry for the previous email. I had to check more. I had a program
of mine eating more than a giga of ram on the server :o(
so no problem with Mysql this time :o)
Regards,
Alex.
-
Before posting, please
Bonjour,
I'm using mysqladmin shutdown to take the server down.
In the past, it was the question of a few seconds. But now, it takes
many minutes to Mysql to shutdown.
I'm wondering if this as something to do with Innobase ... ?
Could there be a reason for Mysqld to take minutes to
Bonjour Heikki,
HT InnoDB will flush the buffer pool and do the purge of old versions
HT at shutdown. It does these operations also when the database is
HT running, but there may be remaining work to do at shutdown.
HT How is the load of your database when you are running it?
HT Does mysqld
Bonjour Warren,
What about InnoDB tables ? Could you change the TYPE to INNODB and
run the same test ? This would be great to compare MYISAM, GEMINI
and INNODB.
Regards,
Alex.
WvdM Goodday to you all
WvdM For those that are interested, I have managed to get my GEMINI tables
Bonjour,
Mysql-3.23.37 with latest patch (patch of the patch from Heikki :)
) under Solaris 2.8 (SPARC 64bit).
I have a table :
CREATE TABLE mybuffer (
id int(10) unsigned NOT NULL auto_increment,
bourse char(3) NOT NULL default '',
service varchar(10) NOT NULL default '',
date
Heikki,
HT can you print the EXPLAIN SELECT ... for the count(*) and the other query?
mysql explain select count(*) from mybuffer where bourse='NAS';
+--+--+---+---+-+---+--+-+
| table| type | possible_keys | key
Bonjour,
Mysql-3.23.37 on Solaris 2.8.
I created this table :
CREATE TABLE livequote_last_PAR (
symbole char(20) NOT NULL default '',
id int(10) unsigned NOT NULL default '0',
date datetime NOT NULL default '-00-00 00:00:00',
last double(17,5) NOT NULL default '0.0',
Heikki,
I did the downgrade to Mysql-3.23.36. I don't have this problem with
3.23.36 !
Alex.
HT check by other queries whether it is the table which contains a
HT duplicate row, or if the SELECT gives a wrong answer.
HT Check what
HT SELECT * FROM ... WHERE symbole='1rPFTE';
HT
Heikki,
Another exemple :
mysql select * from livequote_last_PAR where symbole='1rPFTE' order by id desc;
+-+---+-+--++
| symbole | id| date| last | volume |
+-+---+-+--++
|
Heikki,
HT check by other queries whether it is the table which contains a
HT duplicate row, or if the SELECT gives a wrong answer.
HT Check what
HT SELECT * FROM ... WHERE symbole='1rPFTE';
HT returns.
Really strange :
mysql select * from livequote_last_PAR where symbole='1rPFTE' order
Heikki,
Mysql-3.23.36, Solaris 2.8.
My mysqld crashes when I create a TEMPORARY InnoDB table from the console
if I do a SELECT COUNT(*) before exiting from the console.
No problem with MyISAM tables, no problem with no TEMPORARY table.
mysql CREATE TEMPORARY TABLE mybuffer ( id int
Bonjour,
We found a bug in INNOBASE tables with col_name(length) indexes.
Here is the table :
create table test (
symbole char(20) not null primary key,
libelle char(100) not null,
index id_lib (libelle(1))
) TYPE=INNOBASE;
For the test I insert one row but there can be many,
Bonjour,
Sorry I mistyped the version of Mysql with which we have the bug.
It's Mysql-3.23.36 with patch correcting Solaris and other non Linux
Unix problem.
We found a bug in INNOBASE tables with col_name(length) indexes.
Here is the table :
create table test (
symbole
Heikki,
Could you please tell me (and add in the manual perhaps) if this is
a planned enhancement or not and if it's planned, if it will be
added in a long or a short delay.
Thanks,
Alex.
HT Innobase does not support partial-field indexes. We have to block their
HT usage in the
Bonjour,
I have a table created with :
CREATE TABLE intraday_PAR (
symbole char(20) NOT NULL default '',
date datetime NOT NULL default '-00-00 00:00:00',
first double(17,5) NOT NULL default '0.0',
high double(17,5) NOT NULL default '0.0',
low double(17,5) NOT NULL
Bonjour,
I may have copy all the lines from "mysqladmin process-list".
While I'm doing my "DELETE FROM", another program is doing UPDATE on
the query, but this should not be a problem.
[12:32pm]# mysqladmin proc
Bonjour,
I have re-run my query without the other program trying to UPDATE
the table :
mysql delete from intraday_PAR where date'2001-04-03 00:00:00';
Query OK, 364842 rows affected (32.95 sec)
It works.
Why did both threads lock each other.
There is one doing a lot of UPDATE
Bonjour,
SM Seems like Innobase has introduced some timeout ( I do not see how) on
SM the sleeping connections.
I've been trying a few other things.
When I run Mysql with Innobase, if no client connects, no "Warning
signal 14" messages in the error log.
Then I run a client
Sinisa,
SM There is no thread 4 because may be it has died.
SM If your daemon is multithreaded, then some of it's thread has may be
SM died, thus causing timeout.
SM I hate myself for saying this, as logically there could be no
SM connection, but Just out of curiosity if you build without
Michael,
BAUMEISTER No I did not do "make clean". I didn't knew that I had to in order
BAUMEISTER to update mysqlbug.
MW One must always doe a 'make clean' if one changes any configure arguments.
I'm compiling from a fresh tarball each time.
..
MW I assume you did used gnu make for
Sinisa,
SM No, nothing, No need to test further.
SM Try those Solaris patches, re-install gcc and build from the fresh
SM tarball without debug and with all the options like the last time.
Fresh tarball.
Latest Solaris 2.8 patch applied !
But :
010323 17:38:04 mysqld started
Sinisa,
SM Alex, you ran out of filedescriptors or out of anodes, depending on OS
SM that you are using.
I though filedescriptors could be the problem. I only had 256. I
modified the system in order to have 1024 :
Sinisa,
SM Do you use thread cacheing ??
SM Please turn it off temporarily and see if problem reccurs.
Currently I have :
my.cnf
Sinisa,
SM SM Alex, you ran out of filedescriptors or out of anodes, depending on OS
SM SM that you are using.
SM
SM I though filedescriptors could be the problem. I only had 256. I
SM modified the system in order to have 1024 :
SM
SM
Sinisa,
SM SM Do you use thread cacheing ??
SM
SM SM Please turn it off temporarily and see if problem reccurs.
SM
SM Currently I have :
SM
SM
SM my.cnf
SM
Sinisa,
SM SM SM Alex, you ran out of filedescriptors or out of anodes, depending on
OS
SM SM SM that you are using.
SM SM
SM SM I though filedescriptors could be the problem. I only had 256. I
SM SM modified the system in order to have 1024 :
SM SM
SM SM
Sinisa,
SM Ok, do not erase your binary, save it and try our binary for 2.7.
SM If you still get TCP errors, rebuilding MySQL will be the sole
SM solution.
First, we don't know if it's TCP errors isn't it ?
Second, the binary I'm using now is one I built on my own from the
sources.
Sinisa,
SM Ok, do not erase your binary, save it and try our binary for 2.7.
SM If you still get TCP errors, rebuilding MySQL will be the sole
SM solution.
Another thing to underscore, is that with Mysql-3.23.33 built on my
own from the sources, I never had these messages in the log.
Bonjour Sinisa,
SM BAUMEISTER Alexandre writes:
SM Sinisa,
SM
SM Yes, for now. But in the final application, there will be a lot of
SM connections from network.
SM
SM
SM
SM This entire problem with signal 14 can be due to PHP not using
SM mysql_close.
SM Please try
Sinisa,
SM I've no PHP scripts accessing Mysql for now.
SM
SM Only some little programs written in C which are connecting to Mysql
SM and doing some UPDATE on a few tables.
SM
SM Alex.
SM
SM
SM
SM And you have mysql_close at the end of them ??
These are
Sinisa,
SM Bonjour Sinisa,
SM
SM
SM I've no PHP scripts accessing Mysql for now.
SM
SM Only some little programs written in C which are connecting to Mysql
SM and doing some UPDATE on a few tables.
SM
SM Alex.
SM
SM
SM
SM And you have mysql_close at the
Sinisa,
SM These are daemon programs ... not exiting. I've about 10 programs
SM initialising a connection and then looping, doing some updates on
SM table while receiving data from another input.
SM
SM The schema is :
SM
SM1. Connecting to Mysql
SM2.
Sinisa,
SM If I add a thread on my daemons which executes a mysql_ping() each
SM 30 secondes and increase net_read_timeout to 60 will it be ok ? Does
SM a mysql_ping() reset net_read and net_write timer ?
SM This is all that mysql_ping does (on the server) :
SM send_ok(net);
Sinisa,
SM No, I only referred to bugs as a possibility that daemon exited
SM without closing a connection. That is all.
I'm loggin the activity of my programs (start / stop / statistics
during running). They are not exiting.
SM There were also some reference to problems in Solaris
Sinisa,
SM SM No, I only referred to bugs as a possibility that daemon exited
SM SM without closing a connection. That is all.
SM
SM I'm loggin the activity of my programs (start / stop / statistics
SM during running). They are not exiting.
SM
SM SM There were also some
Bonjour,
SM Ok, start from the fresh tarball. Beside everything, add (in configure
SM options):
SM --with-mit-threads=no
==
My last try (for today) :
CFLAGS=-DSOLARIS \
CXXFLAGS="-DSOLARIS -felide-constructors
Sinisa,
This time I compiled with :
CFLAGS=-DSOLARIS CXXFLAGS="-DSOLARIS -felide-constructors -fno-exceptions -fno-rtti"
./configure
--prefix=/usr/local/mysql --without-bench --enable-thread-safe-client --with-innobase
--without-debug
--with-named-thread-libs=-lthread
Bonjour,
Here is my setup :
Mysql-3.23.35 complied from sources with :
./configure --prefix=/usr/local/mysql --without-bench --enable-thread-safe-client
--with-innobase
System: SunOS
Bonjour Heikki,
HT let us see. A known bug with Innobase is that sys/stat was not included
HT in univ.i in directory mysql/innobase/include. Then Innobase files are
HT created with arbitrary access permissions on some Unixes.
HT Could you look with ls -l what permissions Innobase data file
Bonjour Sinisa,
SM The above could indicate some timeout problem.
SM Can you build mysql --with-debug and start it with --debug and see in
SM the trace file when are the above timeouts reported ??
SM Also, do you have clients that do not call mysql_close at the end ,
SM like PHP programs ?
70 matches
Mail list logo