There are various solutions available for achieving high availability and scalability of MySQL data/service. ‘MySQL Replication’ and ‘MySQL Cluster’ the solutions offered and supported by MySQL. Third-party solutions such as DRBD (Distributed Replicated Block Device) and Heartbeat can also be used or one could implement a combination of these technologies.

The two solutions offered by MySQL are discussed here. Of these, MySQL Cluster is considered a feasible and reliable solution and is our subject of interest.

MySQL Replication:

It enables data from one MySQL server instance to be replicated to another MySQL server instance. It allows replication of data from a single master server to any number of slaves using a simple setup. However, this is an asynchronous replication solution, so the synchronization does not take place in real time, and data replication among all slaves is can not guaranteed always.

• Advantages
  • This can be implemented on any platform (which MySQL supports) and isn’t OS specific.
  • Replication is asynchronous and can be stopped and restarted at any time, and is easily manageable during network issues.
  • Data can be replicated from one master to any number of slaves. This is suitable for environments with heavy reads, but light writes, by spreading the load across multiple slaves.
• Disadvantages
  • Data can only be written to the master.
  • There is no guarantee that data on master and slaves will be consistent at a given point in time. Due to asynchronous replication of data, there may be a small delay between data being written to the master and it being available on the slaves.
• Recommended uses
  • Scale-out solutions that require a large number of reads but fewer writes (for example, web serving).
  • Logging/data analysis of live data. Queries can be performed on the slave without affecting the operation of the master.
  • Online backup (high availability – when used in combination with heartbeat). However, since asynchronous replication is used, the data may be incomplete.
  • Offline backup – Snapshot of existing data at a given point of time can be taken by replicating the data to a slave.

MySQL Cluster:

It is a is a high-availability, high-redundancy synchronous solution that enables multiple MySQL instances to share database information. Data in a cluster can be read from or written to any node within the cluster, and information will be distributed to the other nodes.

This technology enables clustering of in-memory databases in a shared-nothing system. Shared-nothing means each component is allocated its own memory and disk and such an architecture allows the system to work with very inexpensive hardware, and with a minimum of specific requirements for hardware or software. Use of shared storage mechanisms such as SAN, NFS etc is not supported.

• Advantages
  • Offers multiple read and write nodes for data storage.
  • Provides automatic failover between nodes. Only transaction information for the active node being used is lost in the event of a failure.
  • Data on nodes is instantaneously distributed to the other data nodes.
• Disadvantages
  • Available on a limited range of platforms.

Only a development version is available for Microsoft Windows and Apple Mac OSX. It is not supported on SuSE Eneterprise Linux 11 as of this writing. Please check the following link for more details: http://www.mysql.com/support/supportedplatforms/cluster.html

• • Nodes within a cluster should be connected via a LAN; geographically separate nodes are not supported. However, you can replicate from one cluster to another using MySQL Replication, although the replication in this case is still asynchronous.
• Recommended uses
  • Applications that need very high availability, such as telecoms and banking.
  • Applications that require an equal or higher number of writes compared to reads.

There are a various solutions available for achieving high availability and scalability of MySQL data/service. ‘MySQL Replication’ and ‘MySQL Cluster’ the solutions offered and supported by MySQL. Third-party solutions such as DRBD (Distributed Replicated Block Device) and Heartbeat can also be used or one could implement a combination of these technologies.

The two solutions offered by MySQL are discussed here. Of these, MySQL Cluster is considered a feasible and reliable solution and is our subject of interest.

MySQL Replication:

It enables data from one MySQL server instance to be replicated to another MySQL server instance. It allows replication of data from a single master server to any number of slaves using a simple setup. However, this is an asynchronous replication solution, so the synchronization does not take place in real time, and data replication among all slaves is can not guaranteed always.

• Advantages
  • This can be implemented on any platform (which MySQL supports) and isn’t OS specific.
  • Replication is asynchronous and can be stopped and restarted at any time, and is easily manageable during network issues.
  • Data can be replicated from one master to any number of slaves. This is suitable for environments with heavy reads, but light writes, by spreading the load across multiple slaves.
• Disadvantages
  • Data can only be written to the master.
  • There is no guarantee that data on master and slaves will be consistent at a given point in time. Due to asynchronous replication of data, there may be a small delay between data being written to the master and it being available on the slaves.
• Recommended uses
  • Scale-out solutions that require a large number of reads but fewer writes (for example, web serving).
  • Logging/data analysis of live data. Queries can be performed on the slave without affecting the operation of the master.
  • Online backup (high availability – when used in combination with heartbeat). However, since asynchronous replication is used, the data may be incomplete.
  • Offline backup – Snapshot of existing data at a given point of time can be taken by replicating the data to a slave.

MySQL Cluster:

It is a is a high-availability, high-redundancy synchronous solution that enables multiple MySQL instances to share database information. Data in a cluster can be read from or written to any node within the cluster, and information will be distributed to the other nodes.

This technology enables clustering of in-memory databases in a shared-nothing system. Shared-nothing means each component is allocated its own memory and disk and such an architecture allows the system to work with very inexpensive hardware, and with a minimum of specific requirements for hardware or software. Use of shared storage mechanisms such as SAN, NFS etc is not supported.

• Advantages
  • Offers multiple read and write nodes for data storage.
  • Provides automatic failover between nodes. Only transaction information for the active node being used is lost in the event of a failure.
  • Data on nodes is instantaneously distributed to the other data nodes.
• Disadvantages
  • Available on a limited range of platforms.

Only a development version is available for Microsoft Windows and Apple Mac OSX. It is not supported on SuSE Eneterprise Linux 11 as of this writing. Please check the following link for more details: http://www.mysql.com/support/supportedplatforms/cluster.html

• • Nodes within a cluster should be connected via a LAN; geographically separate nodes are not supported. However, you can replicate from one cluster to another using MySQL Replication, although the replication in this case is still asynchronous.
• Recommended uses
  • Applications that need very high availability, such as telecoms and banking.
  • Applications that require an equal or higher number of writes compared to reads.

Issue :

A MySQL database was created and still does exist but it is not being listed in the Plesk interface.

Solution :

This usually happens when the database is not created via Plesk. Instead, it might have been created via SSH.

To get around this, there are two ways :

1. This is the simpler of the two.

Just take a dump of the database.

mysqldump -u <username> -p <password> databasename > databasename.sql

Drop the database.

mysql -u <username> -p <password>
drop database <databasename>

Create the database via Plesk.
Plesk Control Panel >> Domains >> Mysql Database >> Add Database

Restore the database.

mysql databasename < databasename.sql

2. Another method is a bit complex one and is advised for expert users only.

After taking a backup of the current psa database, you will have to edit it from the backend and add the database details in the database table manually.

Issue :

How to reset MySQL ‘root’ password ?

Solution :

-> Stop the mysql daemon.

/etc/init.d/mysqld stop

-> Run the following commands :

mysqld_safe –skip-grant-tables &
mysql -u root
mysql> use mysql;
mysql> update user set password=PASSWORD(“newrootpassword”) where User=’root’;
mysql> flush privileges;
mysql> quit

-> Restart the service.

/etc/init.d/mysqld stop
/etc/init.d/mysqld start

Issue :

While trying to restore a database through phpmyadmin, following error occurs:

#1153 – Got a packet bigger than ‘max_allowed_packet’ bytes

Solution :

This error occurs if the packet size is more than the default value OR the value defined in the my.cnf file.
my.cnf might be blank on a cPanel server by default. You can simply add the variable to it or increase the value if it already exists.

For example,

max_allowed_packet = 2097152

Issue :

After downloading a database backup through cPanel (Download a MySQL database back-up) , the tar file generated is empty!

Solution :

This usually happens when you are logged in to cPanel with the root/reseller password. Log in with the user password and then try to take a back-up. That should work.

Issue :

How to prevent anybody from accessing MySQL databases hosted on the server from any local PC ?

Solution :

This can be achieved in two ways :

1. By blocking the port 3306 . You can use iptables in order to do this :

/sbin/iptables -A INPUT -i eth0 -p tcp –destination-port 3306 -j DROP

OR

2. Put the following in my.cnf :

skip-networking

Source : WHT

Issue :

Following error shows up on the websites which run with mysql.

Warning: session_write_close() [function.session-write-close]: write failed: No
space left on device (28)
in /usr/local/cpanel/base/3rdparty/phpMyAdmin/index.php on line 42

Restarting MySQL fixes the issue temporarily.

Fix :

This usually occurs when /tmp runs out of space. You can check the space by using the following command:

root@server# df -h

Couple of tips to clear some space in /tmp :

1. Check and remove older session files.

root@server# rm -f /tmp/sess*

2. Get rid of semaphores.

root@server# ipcs -s | grep nobody | perl -e ‘while () { @a=split(/\s+/); print `ipcrm sem $a[1]`}’

Issue :

How to login to PHPMyAdmin directly (without logging in to the cPanel interface) ?

Solution :

This can be achieved by the following steps :

First, download and untar the latest version in to your public_html folder.

wget http://downloads.sourceforge.net/project/phpmyadmin/phpMyAdmin/2.11.9.5/phpMyAdmin-2.11.9.5-english.tar.gz
tar -zxvf phpMyAdmin-2.11.9.5-english.tar.gz
cd phpMyAdmin-2.11.9.5-english
mv phpMyAdmin-2.11.9.5-english phpMyAdmin

Then, create a file named config.inc.php and put the following entries in it .

vi config.inc.php
———

$i=0;
$i++;
$cfg['Servers'][$i]['user'] = 'example_test'; //database username
$cfg['Servers'][$i]['password'] = 'qazplm123'; // use here your password
$cfg['Servers'][$i]['auth_type'] = 'config';
?>
———-

Now, you would be able to access it as http://domainname.com/phpmyadmin

For more details on this , please refer this.

Issue:

Cron job required for scheduling automatic checks and repairs on MySQL Databases.

Solution :

Use the following cron :

0 1 * * * /usr/bin/mysqlcheck -Aao –auto-repair -u -p’ ‘

Please note that this will show your MySQL password in the process list. You can get around this by writing the password to ~root/.db_shadow, chowned root, chmod 400, and then reference it as `cat ~/.db_shadow` in the cron command instead of the actual password (note the backslashes in that command) .

Or, if it is not running as root, instead running as a cPanel user, putting the password in a file called .db_shadow in the home of your account, and running the command as follows would achieve the purpose.

/usr/bin/mysqlcheck -Aao –auto-repair -u -p”$(< .db_shadow)"

Issue :

While trying to import a database via PhpMyAdmin , following error occurred :

#1153 – Got a packet bigger than ‘max_allowed_packet’ bytes

Fix:

Get into the file /etc/my.cnf
Change the following to the required value :

================
max_allowed_packet = 16M
================

This should resolve the issue.

Issue :

To list the mysql users and change the password for a particular user via shell.

Solution :

You need to be logged in as ‘root’ in order to do this.

To list the users:

# mysql -u root -p
mysql> SELECT User from mysql.user;

To change the password :

# mysql -u root -p
mysql> SET PASSWORD FOR ‘user’@'hostname’ = PASSWORD(‘newpassword’);
mysql> flush privileges;

Note : When setting database privileges, remember that underscore (_) is a special character, which is like dot (.) in regular expressions.

GRANT ALL ON `db_name`… grants access to many databases, but
GRANT ALL ON `db\_name`… grants access to single database named “db_name”.

Issue :

After upgrading from MySQL4.1 to MySQL5 on a cPanel server, MySQL is failing to start with the error:

ERROR 2002 (HY000): Can’t connect to local MyS QL server through socket ‘/var/lib/mysql/mysql.sock’ (2)

Fix :

To fix this, run the following command:

mv /etc/my.cnf /etc/my.cnf.OLD

Then force update MySQL :

/scripts/mysqlup –force

Running out of disk space is a very common issue.  Be it a web-hosting service provider or the end-user.  I wanted to share a few tips that the technical support staff  at InstaCarma use to resolve disk-space issues on a cPanel server. As we all know, there are various partitions on a server which are meant for different purposes. So, the methods used to deal with space issues on these partitions will also vary. Let us see them one by one.

The /USR partition

1. Restart the httpd service. This might free a little space some times.

2. Check for apache logs like error_log, access_log , suexec_log in /usr/local/apache/logs . These can either be cleared off or if you need the logs then you can take a zipped copy and keep it aside.

3. Same can be done for the files in cPanel logs (/usr/local/cpanel/logs)  as well .

4. Domlogs – Get into  the /usr/local/apache/domlogs/ directory. Run the following command :-

# ls -al -SR | head -10        —> It will list 10 files in the decreasing order according to their size

If the domlog file is too large for a domain then it is possible that awstats is not running . Check whether cpanellogd is running on the server  using  pstree . If not, restart it .
Else, it is possible that awstats for only that particular domain is not updating. Get into the directory /usr/local/cpanel/base and check if any file as ‘awstats.domainname.com.conf’ exists.   If yes , delete that file.

Now,  run /scripts/runweblogs for that user.  It will update the awstats and automatically clear the domlogs file thereafter. Do not delete the domlogs file itself.

5. Remove old and unwanted backups of  ‘apache’  that might have been taken long ago. Also, check for any other duplicate folders that can be removed safely.

6. Remove core files, if any . Normally, some core files (like core.1234) might be present in /usr/local/cpanel/whostmgr/docroot . Check for these and remove them.

—————————————————–

The /HOME partition

1.  Check for unwanted cpmove/tar files for the user accounts in /home and delete them.

2.  Check for any cpmove folders as well in /home

3.  Check the folder ‘cprestore’  in /home for any tar files. Be careful. These should  be deleted only if they are very old,say about 4-5 months.

4. Using find command you can check for unwanted tar files(especially cpmove/backup) in the entire home directory. Check their size. If you think that deleting them will  make considerable difference to the disk space then delete them.

5. Check for core files and delete them.

6.  Search for large accounts and see if they are hosting any prohibited content or violating your company policies. If yes, take appropriate measures as per your policies.

—————————————————–

The /VAR partition

1. Restart the mysql service.

2. Log files in /var/log like exim_mainlog, mailmon.log, messages, mysqld.log, mysql_slow_queries.log  can be deleted after taking backups of the required ones..

Note that there is no point in deleting these log files if the size is very small. Check whether deleting these would make any difference.

If you are deleting them then do re-create the files, or you can use the command ‘echo ” ” > logfile’ to empty the existing files.

3.  Clear Mysql logs like server-name.err or any other  present  in /var/lib/mysql

4. Delete mysql-bin.000* file  in /var/lib/mysql, if present.

5. There might be useless large .sql files   in /var/lib/mysql . Delete them.

6.  Remove unwanted emails from spool. You can use the following code to do that :

exim -bpc
exim -bpru | grep frozen | awk {‘print $3′}|xargs exim -Mrm
exim -bpru | grep “<>” | awk {‘print $3′}|xargs exim -Mrm
exim -bpru | grep “nobody” | awk {‘print $3′}|xargs exim -Mrm
exim -bpru|awk {‘print $3′}|xargs exim -Mrm
exiqgrep -o 86400 -i | xargs exim -Mrm
exim -bpc

7. There might be yum cache files inside the /var/cache folder which can take space. To remove them do:
#  yum clear all
#  yum list all

8. Again, search for any core files ( core dump ) like core.12* etc.

—————————————————–

The /BACKUP partition

1. Remove any tar/cpmove files if they are too old and large.

2. Remove any old  configuration backups ( like apache.bkp) or any log file backup , say 3-4 months old.

3. Check for core files and remove them.

These are some of the general steps that can be followed to deal with space issues on the servers.

Issue :

The phpMyadmin errror “#2002 – The server is not responding (or the local MySQL server’s socket is not correctly configured)”

Fix :

1.First check whether mysql is working in the server or not.

2. Check whether there is a symbolic link from mysql.sock to /tmp. If not, create a symlink between /var/lib/mysql/mysql.sock and /tmp/mysql.sock.
~~~~~~~~~
ln -s /var/lib/mysql/mysql.sock /tmp/mysql.sock
~~~~~~~~~

3. If this doesn’t fix the issue check the file: /usr/local/cPanel/base/3rdparty/phpMyAdmin/config.inc.php and make sure the entries are like this:
~~~~~~~~~~~~~~~~
$cfg['Servers'][$i]['socket'] = ‘/var/lib/mysql/mysql.sock’;
$cfg['Servers'][$i]['connect_type'] = ‘socket’;