Blogging and building live from MySQL Conf 09… If you want more details about what happens here (technical sessions, MySQL 5.4, InnoDB improvements, MySQL Cluster, Sun/Oracle fusion…) or want to enjoy a drink in the Valley, please let me know…
credits : Duncan Davidson
Back to the Dotdeb business : MySQL 5.1.34 has been published with its lot of bugfixes. It is now built and available on Dotdeb for your Lenny/Etch amd64/i386 Debian servers. No change in the packaging, just apt-get upgrade…
Vincent de Lau
After upgrading to 5.1.34 I ran into some trouble with stored procedures. The default config value for thread_stack is to low (128K) to do even empty procedures. Raising it to the default 196K should fix this.
See: http://forums.mysql.com/read.php?98,256564,256564#msg-256564
The message: Thread stack overrun: %ld bytes used of a %ld byte stack, and %ld bytes needed. Use ‘mysqld -O thread_stack=#’ to specify a bigger stack.
Hm
Cannot stop or start after installation on Lenny.
In logs I foudn syntax error 1064
…alter table user add column show_view_priv…
Josh
I upgraded from mysql 5.0 to mysql 5.1 on etch, and I get the same error during post-installation:
ERROR: 1064 You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ‘ALTER T
ABLE user ADD column Show_view_priv enum(‘N’,'Y’) CHARACTER SET utf8 NOT ‘ at line 1
But mysql 5.1 appears to be running fine. I did manually ran mysql_upgrade to be on the safe side. I checked, and the user table already has this column in it. I am just trying to figure out a way to tell aptitude that the package is fully installed so it won’t keep trying. Perhaps there is something els I can do to make things better?
Josh
I think I found the issue, there is a syntax error in the mysql-server-5.1postinst file that I got when I did a apt-get install mysql-server-5.1, when I was running at the time mysql-5.0 (5.0.77). There is a missing comma:
” Show_view_priv=’Y', Create_routine_priv=’Y', Alter_routine_priv=’Y’ “\
You will notice the missing comma at the end. After I fixed that, it seemed to somehow tried to create the plugin table twice. I dropped the table and ran it again and it still gave me a ‘plugin’ table already exists. I just commented that part out. Then it had some sort of issue with the ‘# Automatically added by dh_installinit’ section where it would give a return code of 1. I then decided to run “update-rc.d mysql defaults 19 21 >/dev/null” manually, and then I put in a “exit 0″ right before that section. This allowed it to appear to be fully installed to apt-get. And as far as I can tell, mysql is working perfectly fine.
I am not sure if there really is an issue, or just something I did wrong.
Thanks for providing these wonderful packages over the years.
name
Yes, it is a query bug. Try to change saved maintenace account to root without password.
Raleigh Guevarra
Hi,
I just done with a clean install of LAMP, but when I try to stop or restart the MySQL server this is what I get:
# /etc/init.d/mysql stop
Stopping MySQL database server: mysqld failed!
# /etc/init.d/mysql restart
Stopping MySQL database server: mysqld failed!
Starting MySQL database server: mysqld already running.
Version: ’5.1.34-0.dotdeb.1-log’ socket: ‘/var/run/mysqld/mysqld.sock’ port: 3306 (Debian)
What am I missing? thanks in advance
Guillaume Plessis
@Raleigh : take a look at /var/log/syslog for more details. MySQL seems to be still running. If stuck, kill it : killall -9 mysqld_safe mysqld
Raleigh Guevarra
ok, but do I have to do this, and not able to restart the server normally?
# killall -9 mysqld_safe mysqld
After killing the services and tried to start it, here’s what I got:
#/etc/init.d/mysql start
Starting MySQL database server: mysqld.
Checking for corrupt, not cleanly closed and upgrade needing tables..
EC2-WebHost:~# ERROR 1045 (28000): Access denied for user ‘debian-sys-maint’@'localhost’ (using password: YES)
Thanks in advance
Raleigh Guevarra
Do I still need to run mysqld_safe and not just the mysqld?
# ps aux |grep mysql
root 14975 0.0 0.0 10336 1428 pts/1 S 18:51 0:00 /bin/sh /usr/bin/mysqld_safe
mysql 15101 0.0 0.2 180848 23044 pts/1 Sl 18:51 0:00 /usr/sbin/mysqld –basedir=/usr –datadir=/var/lib/mysql –user=mysql –log-error=/var/lib/mysql/MyDomain.err –pid-file=/var/run/mysqld/mysqld.pid –socket=/var/run/mysqld/mysqld.sock –port=3306
root 15183 0.0 0.0 5292 788 pts/1 R+ 18:53 0:00 grep mysql
Thanks again
Guillaume Plessis
@Raleigh : safe_mysqld is the main binary in charge of launching the mysqld processes. For your error, you’ll have to reset the debian-sys-maint’s one and put it in /etc/mysql/debian.cnf
Raleigh Guevarra
Sorry but how do I reset the debian-sys-maint?
here is the content of /etc/mysql/debian.cnf:
# Automatically generated for Debian scripts. DO NOT TOUCH!
[client]
host = localhost
user = debian-sys-maint
password = *****************
socket = /var/run/mysqld/mysqld.sock
[mysql_upgrade]
host = localhost
user = debian-sys-maint
password = *****************
socket = /var/run/mysqld/mysqld.sock
basedir = /usr
thanks
Guillaume Plessis
@Raleigh : Just change the ********** with a good password (for – bad – example “yourpassword”) then execute this query :
SET PASSWORD FOR ‘debian-sys-maint’@'localhost’ = PASSWORD(‘yourpassword’);
Raleigh Guevarra
should I create user first for ‘debian-sys-maint’@’localhost’? as there is only 1 user, root or should I edit the /etc/mysql/debian.cnf file and change the default password there?
Guillaume Plessis
@Raleigh : You shouldn’t have deleted this user, he’s needed for the Debian administrative tasks. To re-create it, execute this query :
GRANT ALL PRIVILEGES ON *.* TO ‘debian-sys-maint’@'localhost’ IDENTIFIED BY ‘yourpassword’ WITH GRANT OPTION;
Raleigh Guevarra
I didn’t deleted it and also didn’t touched the file, I only learned those things here.
I’m somewhat a newbie with debian and mysql
Ok, I’m going to create a new user debian-sys-maint, is it ok to use the same password in the file debian.cnf or safe to change it and set to mysql?
thanks
Guillaume Plessis
@Raleigh : just do as I said. this use has to be named “debian-sys-maint”.
Raleigh Guevarra
ok done, added user debian-sys-maint on localhost and grant all options, I also changed the password in the file debian.cnf with my own password. did I got that right?
Raleigh Guevarra
after that I stopped the mysql
# /etc/init.d/mysql stop
Stopping MySQL database server: mysqld.
then checked the status
# /etc/init.d/mysql status
MySQL is stopped..
then started it again
# /etc/init.d/mysql start
Starting MySQL database server: mysqld.
Checking for corrupt, not cleanly closed and upgrade needing tables..
does that one ok? not cleanly closed and upgrade needing tables? I have no new database yet other than mysql and information_schema.
thanks
Guillaume Plessis
@Raleigh : All is ok., congratulations!
Raleigh Guevarra
Thank you so much! ^:)^
btw, do you have any documentation about your MySQL debian package?
thanks again
Guillaume Plessis
@Raleigh : take a look at /usr/share/doc/mysql-*/ for Debian-specific documentation.
Jean-Michel
Hi Guillaume,
I just upgraded to mysql-server-5.1 (5.1.34-0.dotdeb.0-log) with Etch on one server and Innodb is not available (parameter have innodb = NO).
On another server that I updated some days ago with the same version (Etch and 5.1.34) it’s ok. Though they have exactly the same my.cnf.
On the net, they say that “have innodb = NO” means that innodb is not compiled with mysql. What can I do (if I can do something …) ?
Guillaume Plessis
@Jean-Michel : all my packages have InnoDB activated (disabling it is insane). Please check that the MySQL related packages are the same on the two servers and that they come from Dotdeb :
dpkg -l ‘*mysql*’ | grep ‘^ii’
Jean-Michel
Yes the packages are the same on both machines, that is :
> dpkg -l ‘*mysql*’ | grep ‘^ii’
ii libdbd-mysql-perl 3.0008-1 A Perl5 database interface to the MySQL data
ii libmysqlclient15off 5.0.81-0.dotdeb.0 MySQL database client library
ii libmysqlclient16 5.1.34-0.dotdeb.0 MySQL database client library
ii mysql-client-5.1 5.1.34-0.dotdeb.0 MySQL database client binaries
ii mysql-common 5.1.34-0.dotdeb.0 MySQL database common files (e.g. /etc/mysql
ii mysql-server-5.1 5.1.34-0.dotdeb.0 MySQL database server binaries
ii php5-mysql 5.2.9-0.dotdeb.1 MySQL module for php5
On the machine which have ‘have Innodb’ = YES, I have an additionnal package :
ii libmysqlclient15-dev 5.0.67-0.dotdeb.1 MySQL database development files
I don’t think if could have an incidence but I have to mention that on the machine with *NO Innodb*, I have got a problem at upgrade with the acpid package with the error :
Paramétrage de acpid (1.0.4-5etch1) …
Starting Advanced Configuration and Power Interface daemon: acpid: can’t open /p roc/acpi/event: No such file or directory
invoke-rc.d: initscript acpid, action “start” failed.
I found some web pages related to this issue but nothing worked for me. But since Mysql is working at least for MyISAM, I don’t think the problems are correlated, are they ?
Jean-Michel
I’m still stuck with my “have innodb NO” problem. Removing and reinstalling mysql-server-5.1 brought nothing new.
Guillaume Plessis
With the same config as yours :
Welcome to the MySQL monitor. Commands end with ; or \g.Your MySQL connection id is 74
Server version: 5.1.34-0.dotdeb.0 (Debian)
Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
mysql> SHOW ENGINES;
+------------+---------+----------------------------------------------------------------+--------------+------+------------+
| Engine | Support | Comment | Transactions | XA | Savepoints |
+------------+---------+----------------------------------------------------------------+--------------+------+------------+
| InnoDB | YES | Supports transactions, row-level locking, and foreign keys | YES | YES | YES |
| MRG_MYISAM | YES | Collection of identical MyISAM tables | NO | NO | NO |
| BLACKHOLE | YES | /dev/null storage engine (anything you write to it disappears) | NO | NO | NO |
| CSV | YES | CSV storage engine | NO | NO | NO |
| MEMORY | YES | Hash based, stored in memory, useful for temporary tables | NO | NO | NO |
| FEDERATED | NO | Federated MySQL storage engine | NULL | NULL | NULL |
| ARCHIVE | YES | Archive storage engine | NO | NO | NO |
| MyISAM | DEFAULT | Default engine as of MySQL 3.23 with great performance | NO | NO | NO |
+------------+---------+----------------------------------------------------------------+--------------+------+------------+
8 rows in set (0.01 sec)
Be sure that “skip-innodb” is commented in /etc/mysql/my.cnf and please restart your MySQL server : /etc/init.d/mysql/restart
Jean-Michel
Here is my “SHOW ENGINES” :
mysql> SHOW ENGINES;+------------+---------+----------------------------------------------------------------+--------------+------+------------+
| Engine | Support | Comment | Transactions | XA | Savepoints |
+------------+---------+----------------------------------------------------------------+--------------+------+------------+
| MyISAM | DEFAULT | Default engine as of MySQL 3.23 with great performance | NO | NO | NO |
| MRG_MYISAM | YES | Collection of identical MyISAM tables | NO | NO | NO |
| BLACKHOLE | YES | /dev/null storage engine (anything you write to it disappears) | NO | NO | NO |
| CSV | YES | CSV storage engine | NO | NO | NO |
| MEMORY | YES | Hash based, stored in memory, useful for temporary tables | NO | NO | NO |
| FEDERATED | NO | Federated MySQL storage engine | NULL | NULL | NULL |
| ARCHIVE | YES | Archive storage engine | NO | NO | NO |
+------------+---------+----------------------------------------------------------------+--------------+------+------------+
7 rows in set (0.00 sec)
The INNODB row is missing.
And yes, I have
#skip-innodb
in my.cnf
On my other server (with the same my.cnf – I took the same file to be sure), I have the 8 rows listed by SHOW ENGINES;
Guillaume Plessis
What about this?
mysql> SELECT VERSION();+-------------------+
| VERSION() |
+-------------------+
| 5.1.34-0.dotdeb.0 |
+-------------------+
1 row in set (0.01 sec)
Jean-Michel
mine is a bit different :
SELECT VERSION();
+———————–+
| VERSION() |
+———————–+
| 5.1.34-0.dotdeb.0-log |
+———————–+
1 row in set (0.00 sec)
Guillaume Plessis
@Jean-Michel : Where does your mysql-server-5.1 package come from? Please show me your sources.list. It seems that it’s a package built “à la Dotdeb”, but not coming from one of the official mirror.
Jean-Michel
here is my source.list :
deb http://mir1.ovh.net/debian/ etch main
deb-src http://mir1.ovh.net/debian/ etch main
deb http://security.debian.org/ etch/updates main contrib
deb-src http://security.debian.org/ etch/updates main contrib
deb http://packages.dotdeb.org etch all
deb-src http://packages.dotdeb.org etch all
Jean-Michel
Thanks Guillaume, you saved me (with some exchanges of emails).
Finally adding some lines to my.cnf solved this issue :
innodb_data_file_path = ibdata1:1G:autoextend
innodb_autoextend_increment=128M
innodb_log_group_home_dir = /home/mysql
innodb_buffer_pool_size = 2G
innodb_additional_mem_pool_size = 20M
# Set .._log_file_size to 25 % of buffer pool size
innodb_log_file_size = 500M
innodb_log_buffer_size = 8M
innodb_flush_log_at_trx_commit = 1
# origin : innodb_lock_wait_timeout = 50 (changed by JMA on 2009-03-06)
innodb_lock_wait_timeout = 100
(take care : my server has 8GB of RAM – the above figures are not for a server with 1GB RAM)
Guillaume Plessis
Jean-Michel’s problem came from his need to change the InnoDB files’ default location. To do this, you have to set BOTH the innodb_data_home_dir and innodb_data_file_path variables.
About the memory settings, these apply only on a 64bits machine too. A 32bits server will never address more than 3GB of RAM (depending on your kernel configuration, even with HIGHMEM) by process. Then, running MySQL on Linux on a amd64 or equivalent machine is a good idea if you have more than 2GB of RAM. The performance boost is noticeable (>=40%).
Daniel Anliker
hi Guillaume,
we have a problem with mysql crashes, is there a way to get support for that ?
we are willing to pay if someone solve our problem …
name
On Debian edit /etc/mysql/debian.cnf:
# Replaces automatically generated for Debian scripts
[client]
host = localhost
user = root
password =
socket = /var/run/mysqld/mysqld.sock
[mysql_upgrade]
host = localhost
user = root
password =
socket = /var/run/mysqld/mysqld.sock
basedir = /usr
name
dotdeb.org made unstable packages
Shame…
Daniel Anliker
@name to you mean me ?
you can see a crash report on
http://forums.mysql.com/read.php?132,264044,264044#msg-264044
Guillaume Plessis
@”name” : Feel free to submit decent bug reports and to give a real name/email address. Please note that replacing “debian-sys-maint” by root in /etc/mysql/debian.cnf (especially with an empty password) is a very bad idea. Don’t be surprised if something goes wrong
@Daniel Anliker : I’ll email you very soon. Waiting this, please gather some information about your system :
- kernel
- debian release
- sources.list
- installed packages and their versions
- suspicious log messages in /var/log/syslog
- a way to reproduce the crash (application, storage engine, connection type…)
Jan Kruis
Hi, i am getting this error on debian lenny
May 28 00:52:51 server-new /etc/init.d/mysql[21093]: 0 processes alive and ‘/usr/bin/mysqladmin –defaults-file=/etc/mysql/debian.cnf ping’ resulted in
May 28 00:52:51 server-new /etc/init.d/mysql[21093]: #007/usr/bin/mysqladmin: connect to server at ‘localhost’ failed
May 28 00:52:51 server-new /etc/init.d/mysql[21093]: error: ‘Can’t connect to local MySQL server through socket ‘/usr/lib/dovecot/var/run/mysqld/mysqld.sock’ (2)’
May 28 00:52:51 server-new /etc/init.d/mysql[21093]: Check that mysqld is running and that the socket: ‘/usr/lib/dovecot/var/run/mysqld/mysqld.sock’ exists!
May 28 00:52:51 server-new /etc/init.d/mysql[21093]:
This is what i have don
btw in this case i have install mysql-cleint and mysql server after putting dotdeb in the /etc/apt/sourcefile.list
* Configuring MySQL
In order to run Postfix and Dovecot chrooted, we need to make some changes to the configuration of MySQL
We are going to store the MySQL socket file inside Dovecot chroot :
#*
mkdir -p /usr/lib/dovecot/var/run/mysqld
cd /usr/lib/dovecot/var/run
chown mysql:root mysqld
#*
We are now symlinking the directory back into /var/run, DSPAM will use mysql socket file and expect it to be located under /var/run/mysqld/mysqld.sock
#*
invoke-rc.d mysql stop
rm -fr /var/run/mysqld
cd /var/run/
ln -s /usr/lib/dovecot/var/run/mysqld/
#*
You should change /etc/mysql/my.cnf and fix the paths :
#*
nano /etc/mysql/my.cnf
#*
[client]
port = 3306
socket = /usr/lib/dovecot/var/run/mysqld/mysqld.sock
[mysqld_safe]
socket = /usr/lib/dovecot/var/run/mysqld/mysqld.sock
nice = 0
[mysqld]
user = mysql
pid-file = /usr/lib/dovecot/var/run/mysqld/mysqld.pid
socket = /usr/lib/dovecot/var/run/mysqld/mysqld.sock
And finally also edit /etc/mysql/debian.cnf and fix the paths :
#*
nano /etc/mysql/debian.cnf
#*
[client]
host = localhost
user = debian-sys-maint
password = xxx
socket = /usr/lib/dovecot/var/run/mysqld/mysqld.sock
[mysql_upgrade]
user = debian-sys-maint
password = xxx
socket = /usr/lib/dovecot/var/run/mysqld/mysqld.sock
basedir = /usr
Restart MySQL :
#*
invoke-rc.d mysql restart
#*
If i install mysql first and change the sourcefie.list whit dotdeb and make de change as above. i get the same error but when i then mark #skip-bdb in the /etc/mysql/my.cnf which is default unmarkt i get is working again
Regrades
Jan Kruis
alikos
Действительно интересно. Ждем новых сообщений на эту тему
Bug innodb mysql 5.1.34
Hi, I’ve installed the dotdeb packages, on xen vm Lenny-amd64 with 4 cpus, 8G Ram, for a lamp server(apache2, php5, mysql and ezpublish). All seems to be ok, until connexions starts. I’ve tested with ab and siege, the innodb thread crash and start automatic recovery. The debian1.err look like this:
090612 9:00:46 [Note] /usr/sbin/mysqld: ready for connections.
Version: ’5.1.34-0.dotdeb.1-log’ socket: ‘/var/run/mysqld/mysqld.sock’ port: 3306 (Debian)
090612 9:01:16 – mysqld got signal 11 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
We will try our best to scrape up some info that will hopefully help diagnose
the problem, but since we have already crashed, something is definitely wrong
and this may fail.
key_buffer_size=16777216
read_buffer_size=131072
max_used_connections=9
max_threads=75
threads_connected=9
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 641148 K
bytes of memory
Hope that’s ok; if not, decrease some variables in the equation.
thd: 0×1717200
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong…
stack_bottom = 0×44209100 thread_stack 0×20000
/usr/sbin/mysqld(my_print_stacktrace+0×29) [0x8abab9]
/usr/sbin/mysqld(handle_segfault+0x34b) [0x5f841b]
/lib/libpthread.so.0 [0x7fb9bfe78a80]
/lib/libunwind.so.7 [0x7fb9be6f6874]
/lib/libunwind.so.7(_ULx86_64_is_signal_frame+0×44) [0x7fb9be6f6dd4]
/lib/libunwind.so.7(_ULx86_64_step+0×36) [0x7fb9be6f7156]
/usr/lib/libtcmalloc.so.0(GetStackTrace(void**, int, int)+0×134) [0x7fb9c00af814]
/usr/lib/libtcmalloc.so.0 [0x7fb9c009d885]
/usr/lib/libtcmalloc.so.0(__libc_malloc+0×205) [0x7fb9c00b1d25]
/usr/sbin/mysqld(my_malloc+0×32) [0x898602]
/usr/sbin/mysqld(ha_innobase::records_in_range(unsigned int, st_key_range*, st_key_range*)+0x3d) [0x779d2d]
/usr/sbin/mysqld [0x6c4b4f]
/usr/sbin/mysqld [0x6c4bb1]
/usr/sbin/mysqld [0x6c4bb1]
/usr/sbin/mysqld [0x6c48c6]
/usr/sbin/mysqld [0x6c52d4]
/usr/sbin/mysqld [0x6c5558]
/usr/sbin/mysqld(SQL_SELECT::test_quick_select(THD*, Bitmap, unsigned long long, unsigned long long, bool)+0x11b7) [0x6ce8f7]
/usr/sbin/mysqld [0x665831]
/usr/sbin/mysqld(JOIN::optimize()+0x5f1) [0x666a31]
/usr/sbin/mysqld(mysql_select(THD*, Item***, TABLE_LIST*, unsigned int, List&, Item*, unsigned int, st_order*, st_order*, Item*, st_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*)+0×98) [0x66a068]
/usr/sbin/mysqld(handle_select(THD*, st_lex*, select_result*, unsigned long)+0x16c) [0x66fc5c]
/usr/sbin/mysqld [0x6044ec]
/usr/sbin/mysqld(mysql_execute_command(THD*)+0x3b60) [0x609930]
/usr/sbin/mysqld(mysql_parse(THD*, char const*, unsigned int, char const**)+0×210) [0x60b800]
/usr/sbin/mysqld(dispatch_command(enum_server_command, THD*, char*, unsigned int)+0×692) [0x60beb2]
/usr/sbin/mysqld(do_command(THD*)+0xe8) [0x60ce38]
/usr/sbin/mysqld(handle_one_connection+0x21e) [0x6007ae]
/lib/libpthread.so.0 [0x7fb9bfe70fc7]
/lib/libc.so.6(clone+0x6d) [0x7fb9be9cc5ad]
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort…
thd->query at 0x7fb92048e410 is an invalid pointer
thd->thread_id=2
thd->killed=NOT_KILLED
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.
090612 09:01:16 mysqld_safe Number of processes running now: 0
090612 09:01:16 mysqld_safe mysqld restarted
InnoDB: The log sequence number in ibdata files does not match
InnoDB: the log sequence number in the ib_logfiles!
090612 9:01:17 InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files…
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer…19 paquets vont être installés, et 2 retirés.
54,5Mo d’espace disque vont être libérés
InnoDB: Last MySQL binlog file position 0 318, file name /var/log/mysql/mysql-bin.000714
090612 9:01:17 InnoDB: Started; log sequence number 0 769649599
090612 9:01:17 [Note] Recovering after a crash using /var/log/mysql/mysql-bin
090612 9:01:17 [Note] Starting crash recovery…
090612 9:01:17 [Note] Crash recovery finished.
090612 9:01:17 [Note] Event Scheduler: Loaded 0 events
090612 9:01:17 [Note] /usr/sbin/mysqld: ready for connections.
Version: ’5.1.34-0.dotdeb.1-log’ socket: ‘/var/run/mysqld/mysqld.sock’ port: 3306 (Debian)
Guillaume Plessis
This issue will be hard to fix, there’s no useful information in the backtrace and your config looks ok.
Are you using something special, like Google Perftools?
Anyway, a 5.1.35 release is on its way. Maybe it will fix this issue.
Doru Bug innodb mysql 5.1.34
Hello,
My name is Doru. Yes, I have activated Google Perftools.
Well, i have to go back to lenny’s default mysql version now, but this one use all memory (8Go) and sometimes crashes the server. The dotdeb version correct the memory leaks (with the same config uses only 4G of 8) but when the number of connexions increases I have the bug.
Thanks for your answer
Guillaume Plessis
Bonjour Doru,
Google Perftools seem to be buggy on amd64. Please disable them and let me know if the problem still occurs.
Doru Bug innodb mysql 5.1.34
Hi,
Indeed, without Google Perftools, it’s working, but I can see the difference. Busy Ram is now ~6G when it was ~4G with the Perftools.
Thank you,
Doru
Guillaume Plessis
@Doru : I’ll update google-perftools really soon, hoping it will fix the amd64 issues.
Doru Bug innodb mysql 5.1.34
Hi, thank you very much. It would be perfect.
Doru
Guillaume Plessis
@Doru : New Google-Perftools 1.3 packages for Lenny have been uploaded. Please tell me if it fixes your issue.
Doru Bug innodb mysql 5.1.34
Hi Guillaume,
I’ve updated with this new google-perftools, but now mysqld don’t start (i’ve inserted the LD_PRELOAD line in /etc/init.d/mysql). The logged error is:
Jun 18 12:07:30 debian1 /etc/init.d/mysql[2436]:
Jun 18 12:08:06 debian1 kernel: [593293.212217] mysqld_safe[2472]: segfault at 20 ip 7fe5beab9216 sp 7fffc6f1c560 error 4 in libtcmalloc.so.0.0.0[7fe5bea99000+3c000]
and the script error answer:
/etc/init.d/mysql start
Starting MySQL database server: mysqld/etc/init.d/mysql: line 108: 2472 Erreur de segmentation /usr/bin/mysqld_safe > /dev/null 2>&1
I don’t knok if this helps you, I hope so.
Doru
David Yan
Hi, 5.1.34-0.dotdeb.1 just crashed on my amd64 server. http://dev.mysql.com/doc/refman/5.1/en/news-5-1-35.html shows it fixes a number of crashes. When can we expect a 5.1.35 dotdeb package? Thanks!
090620 8:52:30 – mysqld got signal 11 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
We will try our best to scrape up some info that will hopefully help diagnose
the problem, but since we have already crashed, something is definitely wrong
and this may fail.
key_buffer_size=536870912
read_buffer_size=131072
max_used_connections=709
max_threads=1000
threads_connected=31
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 4758475 K
bytes of memory
Hope that’s ok; if not, decrease some variables in the equation.
thd: 0x7f039c0f5f30
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong…
stack_bottom = 0x7f039ac580d0 thread_stack 0×80000
/usr/sbin/mysqld(my_print_stacktrace+0×29) [0x8abab9]
/usr/sbin/mysqld(handle_segfault+0x34b) [0x5f841b]
/lib/libpthread.so.0 [0x7f0b16bc7080]
/usr/sbin/mysqld(lock_sec_rec_cons_read_sees+0×25) [0x7c8345]
/usr/sbin/mysqld(row_search_for_mysql+0×1920) [0x8092d0]
/usr/sbin/mysqld(ha_innobase::general_fetch(unsigned char*, unsigned int, unsigned int)+0x7c) [0x7784fc]
/usr/sbin/mysqld [0x64fc3a]
/usr/sbin/mysqld(sub_select(JOIN*, st_join_table*, bool)+0×89) [0x6562c9]
/usr/sbin/mysqld [0x650185]
/usr/sbin/mysqld(sub_select(JOIN*, st_join_table*, bool)+0×78) [0x6562b8]
/usr/sbin/mysqld [0x6592fd]
/usr/sbin/mysqld(JOIN::exec()+0x51d) [0x66de5d]
/usr/sbin/mysqld(mysql_select(THD*, Item***, TABLE_LIST*, unsigned int, List&, Item*, unsigned int, st_order*, st_order*, Item*, st_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*)+0×124) [0x66a0f4]
/usr/sbin/mysqld(handle_select(THD*, st_lex*, select_result*, unsigned long)+0x16c) [0x66fc5c]
/usr/sbin/mysqld [0x6044ec]
/usr/sbin/mysqld(mysql_execute_command(THD*)+0x3b60) [0x609930]
/usr/sbin/mysqld(mysql_parse(THD*, char const*, unsigned int, char const**)+0×210) [0x60b800]
/usr/sbin/mysqld(dispatch_command(enum_server_command, THD*, char*, unsigned int)+0×692) [0x60beb2]
/usr/sbin/mysqld(do_command(THD*)+0xe8) [0x60ce38]
/usr/sbin/mysqld(handle_one_connection+0x21e) [0x6007ae]
/lib/libpthread.so.0 [0x7f0b16bbf3ba]
/lib/libc.so.6(clone+0x6d) [0x7f0b15708fcd]
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort…
thd->query at 0x7f03a85be190 is an invalid pointer
thd->thread_id=10637885
thd->killed=NOT_KILLED
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.
pure virtual method called
terminate called without an active exception
Fatal signal 6 while backtracing
Guillaume Plessis
@David Yan : Please desactivate the Google Perftools, they crash the MySQL server.
David Yan
I’m not using Google Perftools…
The BLION Corp.
Hello,
After having upgraded Mysql, service began and crash… Sometimes, Mysql just does not live more than few seconds.
I went to 5.1.35 provided by Mysql (binary package), deleted everything, and recreate everything. I don’t have any crash yet. We’ll see wether they come backup or not.
I don’t use libgoogle…
For information, this is a big mysql server
# mysqladmin status ; sleep 60 ; mysqladmin status
Uptime: 134474 Threads: 37 Questions: 490612571 Slow queries: 13068 Opens: 26014 Flush tables: 1 Open tables: 24925 Queries per second avg: 3648.382
Uptime: 134534 Threads: 19 Questions: 491088014 Slow queries: 13068 Opens: 26014 Flush tables: 1 Open tables: 24925 Queries per second avg: 3650.289
# echo $(( (491088014 – 490612571) / 60 ))
7924
http://bugs.mysql.com/bug.php?id=45489
Sebastian Harnau
MySQL 5.1.36 is online on MySQL-Website
Guillaume Plessis
@Sebastian Harnau : MySQL 5.1.36 has been uploaded on Dotdeb for Etch and Lenny. A news is coming.
David Yan
@Sebastian Harnau: yeah, i’m not using dotdeb any more… too slow… sorry.
Guillaume Plessis
@Sebastian Harnau : I don’t know why the official Debian maintainer does not update the mysql-dfsg-5.1 packages (my packages are just backports from experimental to stable) : http://packages.qa.debian.org/m/mysql-dfsg-5.1.html
I’ll backport a newer release as soon as it’s available on Debian experimental
@David Yan : No problem. If you have a “faster” repository than Debian experimental with clean Debian packages, let me know.
Sebastian Harnau
@Guillaume Plessis : Thanks for information, now it is clear, why it takes some time to provide these packages. And of course: Thanks for your work!!!
Guillaume Plessis
MySQL 5.1.36 packages are being worked on : http://twitter.com/ntretkow/status/2538928814
Qwertz
Hi,
after apt-get installation i get a error in phpMyAdmin: “Your PHP MySQL library version 5.0.51a differs from your MySQL server version 5.1.34. This may cause unpredictable behavior.”.
php5-mysql and libapache2-mod-php5 are allready “up to date”.
What should i do?
Tanks.
Guillaume Plessis
@Qwertz : as said in many comments on this site, that’s just a warning and there’s no incompatibility between 5.0 libraries and a 5.1 server in the PHPMyAdmin case. It should be fixed in PHP 5.3 with the inclusion of the MySQL native driver.
Try to comment this warning in PHPMyAdmin (main.php).
Sebastian Harnau
@Guillaume Plessis : Thanks for information!