I have a AWS RDS PostgreSQL 13 server with some databases. I have to create an empty copy of one database (empty means schema (tables, views, functions) + security (users, roles)).
Is pg_dump -s what I am looking for?
Thanks!
pg_dump -d db_name -s. You will also need to do pg_dumpall -g to get the global data e.g. roles. This will get all global data for the Postgres cluster, so you may have more then you need for the particular database.
Postgres allows the use of any existing database on the server as a template when creating a new database. I'm not sure whether pgAdmin gives you the option on the create database dialog but you should be able to execute the following in a query window if it doesn't:
CREATE DATABASE newdb WITH TEMPLATE originaldb OWNER dbuser;
Still, you may get:
ERROR: source database "originaldb" is being accessed by other users
To disconnect all other users from the database, you can use this query:
SELECT pg_terminate_backend(pg_stat_activity.pid) FROM pg_stat_activity WHERE pg_stat_activity.datname = 'originaldb' AND pid <> pg_backend_pid();
Related
Is there a simple way to create a copy of a database or schema in PostgreSQL 8.1?
I'm testing some software which does a lot of updates to a particular schema within a database, and I'd like to make a copy of it so I can run some comparisons against the original.
If it's on the same server, you just use the CREATE DATABASE command with the TEMPLATE parameter. For example:
CREATE DATABASE newdb WITH TEMPLATE olddb;
pg_dump with the --schema-only option.
If you have to copy the schema from the local database to a remote database, you may use one of the following two options.
Option A
Copy the schema from the local database to a dump file.
pg_dump -U postgres -Cs database > dump_file
Copy the dump file from the local server to the remote server.
scp localuser#localhost:dump_file remoteuser#remotehost:dump_file
Connect to the remote server.
ssh remoteuser#remotehost
Copy the schema from the dump file to the remote database.
psql -U postgres database < dump_file
Option B
Copy the schema directly from the local database to the remote database without using an intermediate file.
pg_dump -h localhost -U postgres -Cs database | psql -h remotehost -U postgres database
This blog post might prove helpful for you if you want to learn more about options for copying the database using pg_dump.
This can be done by running the following command:
CREATE DATABASE [Database to create] WITH TEMPLATE [Database to copy] OWNER [Your username];
Once filled in with your database names and your username, this will create a copy of the specified database. This will work as long as there are no other active connections to the database you wish to copy. If there are other active connections you can temporarily terminate the connections by using this command first:
SELECT pg_terminate_backend(pg_stat_activity.pid)
FROM pg_stat_activity
WHERE pg_stat_activity.datname = '[Database to copy]'
AND pid <> pg_backend_pid();
A good article that I wrote for Chartio's Data School which goes a bit more in depth on how to do this can be found here:
https://dataschool.com/learn/how-to-create-a-copy-of-a-database-in-postgresql-using-psql
How can I create a postgres user who has admin access only to one database but cannot inspect or interfere with other databases in the postgres instance? The use case is I'm creating a multitenant application
where each tenant gets his own database in the postgresql instance and can create schemas, tables etc and
perhaps use a few pg_tables to inspect his own database but not others.
cannot change the name of the database as it's controlled by me
EDIT: Added more constraints
That's fairly trivial:
CREATE DATABASE newdb;
GRANT CREATE ON DATABASE newdb TO newdba;
Add pg_hba.conf entries to allow newdba to connect to newdb only.
I would like to copy data from table A database A to table b database b in postgres.
I am able to use copy to write data out to a file from table A in DB A
but when I try copy into database b table B it says "ERROR: must be super user to copy to and from file"
Please help and let me know how I should data from a flatfile into a table in a database.... the db is postgres 9.x
According to the documentation:
COPY naming a file or command is only allowed to database superusers or users who are granted one of the default roles pg_read_server_files, pg_write_server_files, or pg_execute_server_program, since it allows reading or writing any file or running a program that the server has privileges to access.
However you can use pipeline to transmit data from one database to another one without using of intermediate file(s):
psql -d A -c "copy a to stdout" | psql -d B -c "copy b from stdin"
Read more about psql utility and its usage.
I'm trying to rename a database in my Redshift cluster.
You cannot rename the database when you're connected to it so I've created a temporary database, reconnected with SQL Workbench to the temporary db and issued:
ALTER DATABASE olddb RENAME to newdb;
I get an error stating ERROR: database "olddb" is being accessed by other users [SQL State=55006]
I've checked who is connected and there appear to be some connections from user rdsdb to the database. I assume this is a service account that AWS Redshift use to perform maintenance tasks etc.
How can I rename the database when this superuser is connected?
Many thanks.
You cannot alter the name of (or delete!) the database that is created during the initial cluster creation. I don't believe this is mentioned in the docs but I've confirmed it with them.
We can change the database name which is already created.
Detailed steps on how to do
Connect to the old database and create a new database if you do not have another one already.
create database databasebasename
In this example, I will call the databasename as 'newdb'.
Connect to newdb using connecting string as, jdbc:redshift://.us-east-1.redshift.amazonaws.com:8192/newdb, with the same password and username of your superuser (or the other eligible users as mentioned above).
Now you can alter the database name. Substitute 'database_name_new' with the desired databasename.
alter database old-db-name rename to database_name_new;
If there are any active sessions, you'll have to kill them. To find the pid of active sessions:
select * from STV_SESSIONS where user_name='rdsdb';
Then to kill a session:
SELECT
pg_terminate_backend(<pid>)
FROM
pg_stat_activity
WHERE
-- don't kill my own connection!
procpid <> pg_backend_pid()
-- don't kill the connections to other databases
AND datname = '<old-db-name>';
Once complete, you can connect back to that new database using the new name in the connection string as
jdbc:redshift://<cluser-id>.us-east-1.redshift.amazonaws.com:8192/database_name_new
You can delete the temporary 'newdb'.
drop database databasebasename
That's possible now -- I just renamed the database that was created during the initial cluster creation.
We had a similar situation.
Step 1: Connect to the database which is not the one you are trying to rename. Check the same by executing SELECT CURRENT_DATABASE();.
Step 2: Execute the query below -
SELECT
ss.*, 'select pg_terminate_backend('||process||');'
FROM
stv_sessions ss
ORDER BY
db_name;
The output of the query will have a column at the end with the select statements. Execute those to kill the sessions.
Step 3(Optional): If you are not the owner of the database try to modify the ownership of the database -
ALTER DATABASE <database to be renamed>
OWNER TO <user which is going to do the rename>;
Step 4: Rename the database
I am looking for a basic script/command that will create a copy of a live database (let name them mydb and mydb_test, both on the same server).
Requirements
it has to run even if the mydb_test already exists and have records
it has to work even if mydb and mydb_test do have existing connections
it have to clean the potentially existing database if necessary
Hints:
drop database cannot be used if you have existing connections
The simplest and fastest method to create a complete copy of an existing (live) database is to use CREATE DATABASE with a TEMPLATE:
CREATE DATABASE mydb_test TEMPLATE mydb;
However, there is an important limitation violating your second requirement: the template (source) database cannot have additional connections to it. I quote the manual:
It is possible to create additional template databases, and indeed one
can copy any database in a cluster by specifying its name as the
template for CREATE DATABASE. It is important to understand, however,
that this is not (yet) intended as a general-purpose "COPY DATABASE"
facility. The principal limitation is that no other sessions can be
connected to the source database while it is being copied. CREATE DATABASE
will fail if any other connection exists when it starts; during
the copy operation, new connections to the source database are prevented.
You can terminate all sessions to the template database if you have the necessary privileges with pg_terminate_backend().
To temporarily disallow reconnects, revoke the CONNECT privilege (and GRANT back later).
REVOKE CONNECT ON DATABASE mydb FROM PUBLIC;
-- while connected to another DB - like the default maintenance DB "postgres"
SELECT pg_terminate_backend(pid)
FROM pg_stat_activity
WHERE datname = 'mydb' -- name of prospective template db
AND pid <> pg_backend_pid(); -- don't kill your own session
CREATE DATABASE mydb_test TEMPLATE mydb;
GRANT CONNECT ON DATABASE mydb TO PUBLIC; -- only if they had it before
In versions before Postgres 9.2 use procpid instead of pid:
How to drop a PostgreSQL database if there are active connections to it?
Related:
Force drop db while others may be connected
If you cannot afford to terminate concurrent sessions, go with piping the output of pg_dump to psql like has been suggested by other answers already.
That's what I was looking for, but I had to compile it myself :P
I only wish I knew a way to keep the same user and not having to put it inside the script.
#!/bin/bash
DB_SRC=conf
DB_DST=conf_test
DB_OWNER=confuser
T="$(date +%s)"
psql -c "select pg_terminate_backend(procpid) from pg_stat_activity where datname='$DB_DST';" || { echo "disconnect users failed"; exit 1; }
psql -c "drop database if exists $DB_DST;" || { echo "drop failed"; exit 1; }
psql -c "create database $DB_DST owner confuser;" || { echo "create failed"; exit 1; }
pg_dump $DB_SRC|psql $DB_DST || { echo "dump/restore failed"; exit 1; }
T="$(($(date +%s)-T))"
echo "Time in seconds: ${T}"
Since you didn't say it was a problem to drop objects in the database, I think running pg_dump with the --clean option will do what you want. You can pipe the output of pg_dump into psql for this sort of thing.
Are you looking into the Hot Standby with Streaming Replication here?