I am fairly new with PQSQL and am slowly picking things up - I have added a new disk and would like to do two things:
Restore a backup to this new disk - /hda2/pgdata/
Move a database from /hda1/pgdata to /hda2/pgdata/
Q1. Use pg_restore to restore a database. Check out the documentation which is very clear.
One important thing to remember, if you want to move to a later version of PostgreSQL use the later version of pg_dump to create a backup dump file. For example if you want to move from PostgreSQL version 8.3 to version 8.4, then create a backup dump file using pg_dump from version 8.4 and then use pg_restore 8.4 to recreate database in the 8.4 server.
http://www.postgresql.org/docs/8.4/static/app-pgrestore.html
Q2. Backup and restore is a safe way of doing it. Before restoring one can create a tablespace on the new disk and place the database in that space.
CREATE DATABASE mydb TABLESPACE myspace;
http://www.postgresql.org/docs/8.4/interactive/manage-ag-tablespaces.html
Simple command to restore database
Open PSQL Command console
Provide credentials
go to specific database that you need to restore (If the dataabse is not there Create empty database
/i < sql Dump file Path > e.g. \i /usr/local/pgsql/db20121109.sql
Related
I have taken full backup of postgresql database which consists of 100 databases. The backup format is .sql (eg pg_dumpall.exe -U postgres > D:\Backup\fullbkp.sql) now one of my database got crashed and I want to extract this file to get that database backup only for restoration.
I have searched a lot but couldn't find any way to decompress so that I can get that particular database from full backup file.
Please suggest !!!!
Regards
Sadam
Such a backup is not compressed. Also, it contains a backup of all databases in th cluster, and there is no easy way to extract a single database.
Create a new PostgreSQL cluster with initdb, restore the dump there using psql, then use pg_dump to extract the single database you need.
My last working database back up of an Odoo13CE system was a full one, including the file store. I'm getting timeouts when trying to restore "a copy" via Odoo database manager page. Thought I could just do a partial restore (dump.sql & manifest.json), dump the filestore, recompress and upload and that brought everything down to its knees (Errored w/" no *.dump file found). So logged into server and dropped my failed restore and restarted odoo service and all is back to somewhat normal, with the database I want to replace active.
Is there a way to convert that .sql to a .dump or some other way to get my .sql to be added to my pgdb? I'm fairly green re: psql so if I'm missing something simple, please feel free to shove it down my throat.
TIA
to restore sql back up file to a new database:
psql YOUR_DATABASE_NAME < YOUR_FILENAME
You can read more about restoring/back up Postgres Db here: https://www.postgresql.org/docs/11/backup-dump.html
Restoring the heavy size database(with file store) you have to increase the limit of the server to continue your process.
Add the parameter on your path
--limit-time-cpu=6000 --limit-time-real=12000
Restore the SQL File
psql database_name < your_file.sql
Restore the Dump File
pg_restore -d database_name < your_file.dump
It is know fact that backup and restore is a slow in Postgres
I'd like to deploy database to PostgreSQL server as fast as posible (Like it is possible in MS SQL just file copy and attach).
So If I:
backup and restore schema only.
And than copy database oid folder (data files) in to the appropriate oid ?
Will it work?
of not what I also need to be consider.
No, that won't work.
A database backup and restore can never be faster than copying the files is. So stop the database, copy the complete cluster and start PostgreSQL again. It won't get any faster.
I'm using pgAdmin to restore PostgreSQL database. To restore the database I need to delete, drop and remake it. How to restore the database without deleting and remaking it?
This cannot be done in pgAdmin or with any database tools. Regular backup files cannot be restored without deleting the data first because they consist of normal COPY statements which will fail if you have rows in the database (primary keys collide etc).
For a simple way to get back to an earlier snapshot in a testing environment take a look at PostgreSQL documentation - 24.2. File System Level Backup:
For backup:
Shut down your database
copy all the files from your data directory
For restore:
Shut down your database
replace your data directory with the backup directory
Note:
the size of the data might be significantly larger than with a regular backup especially if you have a lot of indexes
this is a server wide backup so you can't do this on individual databases
don't attempt to use it on a different version of PostgreSQL
this really deletes the data too - by replacing it with the backup
Also with regular backups you don't have to do a DROP TABLE if you do a data-only restore with pg_restore --data-only for example. You still have to delete the data though.
I want to make a copy of a live Firebird .fdb database.
I understand that simply copying it could result in a corrupted database and I looked into using the gbak command because it is able to perform a backup while the database is running.
So that will give me a database backup but I would need to restore this before I can use it. My database is nearly 1GB and takes 10 minutes to restore which is too long. Is there any other method to simply copy a Firebird live database from one location to another safely?
So far I have used the following to backup (which works):
gbak -v -t -user SYSDBA -password "masterkey" 127.0.0.1:"C:/Files/Live/Database.fdb" "C:\Test\Test.fbk"
I also tried using the following to backup and restore at the same time:
gbak -c [options] <source database> stdout | gbak -r [options] stdin <target database>
but this kept giving the error:
Done with volume #1, "new.gbak"
Press return to reopen that file, or type a new
name followed by return to open a different file.
The risk of corruption is due to the way Firebird writes it file. Your copy might contain inconsistent data when copied at the same time that Firebird (re)writes a datapage. As far as I know the only real risk of corruption is during writes of index pages (and then only for index page splits), otherwise it just leads to inconsistent data and dangling transactions (which wouldn't be visible anyway as the transaction is not committed).
If you really don't want to use a backup, you can set the Firebird database in a backup state. This freezes the database and writes the changes to a delta file instead. You enable this state using ALTER DATABASE BEGIN BACKUP and end this state using ALTER DATABASE END BACKUP. See the documentation for ALTER DATABASE. This command was added in Firebird 2.0.
For the second part of your question (which really should have been posted as a separate question):
The command gbak -c doesn't create a backup, it Creates a database from a backup (it is the brother of -r (replace) but doesn't overwrite an existing database. See Restore Switches for more information.
See Create a Database Clone Without a Dump File for an example of how you can make a backup like this:
gbak -backup emptest stdout | gbak -replace stdin emptest_2
You can do
alter database begin backup;
then copy the file using standard file copy of your OS and
alter database end backup;
Also I strongly recommend reading these pages [1] [2] about nbackup.