Installing and Configuring PostgreSQL in Ubuntu

The installation command in Ubuntu:

sudo apt-get install postgresql postgresql-client postgresql-contrib

You can install a graphical client for easy management:

sudo apt-get install pgadmin3

or

sudo apt-get install phppgadmin

To access from outside, open the configuration file (Ctrl + X to exit):

sudo nano /etc/postgresql/9.3/main/postgresql.conf

And uncomment the line:

listen_addresses = 'localhost'

In the pg_hba.conf file, we specify which addresses are allowed access:

host all all 192.168.0.1/32 md5

Restart PostgreSQL so that the changes to the configuration files take effect:

sudo service postgresql restart

Here is the password for the postgres user:

sudo -u postgres psql
ALTER USER postgres with encrypted password 'PASSWORD';
CTRL+Z

An example of testing a connection from a remote computer:

psql -h СЕРВЕР -U postgres -W

If PostgreSQL is to be used with Apache2, then we’ll install the other components:

sudo apt-get install apache2 libapache2-mod-php5 php5 php5-common php5-gd php5-pgsql

Solution of the error “Using unique option prefix pass instead of password is deprecated …”

I noticed some errors coming to the root mail with the subject and the text:

Cron /usr/bin/test -x /usr/local/cpanel/scripts/update_db_cache && /usr/local/cpanel/scripts/update_db_cache
Warning: Using unique option prefix pass instead of password is deprecated and will be removed in a future release. Please use the full name instead.

The problem lies in the file ~ / .my.cnf, that is /root/.my.cnf in my case.

In which the parameter “pass” is obsolete and must be changed to a new “password”.

For example, that’s how it was when the error occurred:

[client]
user=root
pass=password

Changed to password and the error disappeared:

[client]
user=root
password=password

See also:
Connecting to MySQL from localhost without entering a password

How to fix error “Table ‘name’ is marked as crashed and last (automatic?) repair failed”

Once in the FreeRADIUS logs I noticed a MySQL error:

Table ‘./radius/radacct’ is marked as crashed and last (automatic?) repair failed

As it turned out, the radacct table was damaged, since the data there were not particularly important, then the entire table was cleaned.
You can clean up via phpMyAdmin or SQL query:

truncate table TableName

A bit later for the experiment I decided to break the whole database, took another large table in general from another application, about 8 gigabytes in size and 80 million lines.
I applied to it SQL query to clean up old rows before the date specified in the query and rebooted at that moment MySQL, the request was interrupted, the database was left intact, executed the request to optimize the database and again rebooted MySQL, eventually got a corrupted database and a similar error:

#144 – Table ‘name’ is marked as crashed and last (automatic?) repair failed

To restore the database, you must stop the MySQL server (if the table is not used, then you can not stop it):

sudo service mysql stop

Let’s move to the directory with the database:

cd /var/lib/mysql/$DATABASE_NAME

Execute the command to restore the specified table:

myisamchk -r -o -f -v $TABLE_NAME

Upon completion, if you stopped the MySQL server, then run it:

sudo service mysql start

Similarly, on the test, also to speed up the process, the table was restored by copying it to another more powerful server, namely three files /var/lib/mysql/$DATABASE_NAME/ ($TABLE_NAME.MYD, $TABLE_NAME.MYI, $TABLE_NAME.frm).

How to change MySQL database encoding and its tables

Here is an example of changing the MySQL encoding of a database and tables.
Before any actions on important data it is necessary to make a backup copy, for example:

mysqldump -u USER -h localhost -p BASE | gzip -c > backup_base_`date +%Y-%m-%d`.sql.gz

For the test, we will connect to MySQL and create a couple of new databases without specifying the encoding and specifying:

mysql -u root -p
CREATE DATABASE test_db1;
CREATE DATABASE test_db2 CHARACTER SET utf8 COLLATE utf8_general_ci;

Create a test table in the first database and see its encoding:

USE test_db1;

CREATE TABLE users (
id INT(6) UNSIGNED AUTO_INCREMENT PRIMARY KEY,
firstname VARCHAR(30) NOT NULL,
lastname VARCHAR(30) NOT NULL,
email VARCHAR(50),
reg_date TIMESTAMP
);

show table status like 'users';

Create a test table in the second database and see its encoding:

USE test_db2;

CREATE TABLE users (
id INT(6) UNSIGNED AUTO_INCREMENT PRIMARY KEY,
firstname VARCHAR(30) NOT NULL,
lastname VARCHAR(30) NOT NULL,
email VARCHAR(50),
reg_date TIMESTAMP
);

show table status;

Let’s also look at the encoding of both databases:

SELECT default_character_set_name FROM information_schema.SCHEMATA WHERE schema_name = "test_db1";
SELECT default_character_set_name FROM information_schema.SCHEMATA WHERE schema_name = "test_db2";

To see the encoding of a column in a specific table, you can do this:

SELECT character_set_name FROM information_schema.`COLUMNS`
WHERE table_schema = "test_db1"
AND table_name = "users"
AND column_name = "firstname";

In my case, the table in the first database was encoded with latin1_swedish_ci, since it is standard, and in the second one utf8_general_ci, since I specified it beforehand.

You can see the table of possible encodings by such requests:

show collation;
show collation like 'utf8%';
show collation like 'latin1%';

View existing databases as follows:

show databases;

View existing tables in the database:

USE test_db1;
show tables;

Now we change the encoding of the first database and its tables to utf8 and immediately check:

ALTER DATABASE `test_db1` DEFAULT CHARACTER SET utf8 COLLATE utf8_general_ci;
USE test_db1;
ALTER TABLE `test_db1`.`users` CONVERT TO CHARACTER SET utf8 COLLATE utf8_general_ci;
show table status;

If you need to change the encoding in a sql file, open it in the Notepad ++ editor, for example, convert it to UTF-8/without BOM, and if the encoding in SET NAMES is specified at the beginning of the file, change it there, then you can import the file into the database.

How to create a MySQL user and configure access rights

To create a user, we first connect to the MySQL server console:

mysql

Let’s see what users are:

select * from mysql.user;
select user,host from mysql.user;

Create a user (where localhost is specified from where the user can connect, you can specify the IP address, localhost – from the local machine where the MySQL server itself, or % from any addresses):

CREATE USER 'user'@'localhost' IDENTIFIED BY 'password';

If you intend to connect not only locally, you need to comment out the line in my.cnf:

#bind-address = 127.0.0.1

And restart the MySQL server:

sudo service mysql restart

After that, I recommend restricting access to MySQL using IPTables.
See also – Configuring IPTables

To assign the newly created user unlimited permissions to a specific database, execute the following command:

GRANT ALL PRIVILEGES ON database_name.* TO 'user'@'localhost';

If necessary on all bases:

GRANT ALL PRIVILEGES ON *.* TO 'user'@'localhost';

You can specify specific access rights:

GRANT SELECT ON database_name.* TO 'user'@'localhost';
GRANT SELECT, INSERT ON database_name.table_name TO user@192.168.1.5;

If you want to create a new database:

CREATE DATABASE database_name;

For the changes to take effect, execute:

FLUSH PRIVILEGES;

You can delete the user as follows:

DROP USER 'user'@'localhost';

Example of viewing privileges:

SHOW GRANTS FOR 'user'@'localhost';
SHOW GRANTS;
SELECT * FROM information_schema.user_privileges;

The solution of error “ERROR 1067 (42000) at line 211: Invalid default value for ‘blablabla'”

I noticed once when importing a sql file the following error:

ERROR 1067 (42000) at line 211: Invalid default value for ‘blablabla’

It arises because new versions of MySQL server use strict mode and parameters such as NO_ZERO_DATE do not allow entering date values like ‘0000-00-00’ into the database.

Connect to mysql server:

mysql -u root -p

Execute a query that displays the values of sql_mode:

show variables like 'sql_mode';

Copy the string with these values and exit mysql:

exit

Open the configuration file for example in the text editor nano (Ctrl+X for exit, y/n for saving or canceling the changes):

sudo nano /etc/mysql/mysql.conf.d/mysqld.cnf

I did not have sql_mode = in the file, so at the end of the file I inserted the line with the previously copied values, removing NO_ZERO_IN_DATE, NO_ZERO_DATE from it, in my case, the following happened:

sql_mode=ONLY_FULL_GROUP_BY,STRICT_TRANS_TABLES,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION

Restart mysql to apply the changes:

sudo service mysql restart

Done, now when importing this error should not be.