Günce Sarıkaya

Hi all, If you have lots of user and table/schema on your database you can not give permission per table but you can follow this steps. It's gonna be manual process but sometimes you can not give whole privileges on entire schema in your databases to users/roles. So, in that case you should generate your scripts. GRANT SELECT ON ALL TABLES IN SCHEMA schema_name TO your_role; GRANT USAGE ON SCHEMA schema_name TO your_role; GRANT EXECUTE ON ALL FUNCTIONS IN SCHEMA schema_name TO your_role; If you don't want to give all privileges and make specific for users also you can generate a sql script for whole schema for user or your all users like; Generate GRANT script for granting privileges whole schemas for a role; SELECT DISTINCT 'GRANT SELECT ON ALL TABLES IN SCHEMA '|| schemaname || 'TO user_role;' FROM pg_stat_user_tables; Generate USAGE script for whole schemas; SELECT DISTINCT 'GRANT USAGE ON SCHEMA '|| schemaname || 'TO user_role;...

Merhaba, PostgreSQL 10 ile gelen yeniliklerden biri olan Logical Replication'dan bahsetmek isterim. Logical Replication (LR), veri objelerini ve üzerlerindeki değişiklikleri, replikasyon kimliklerine göre (primary key gibi) replike eder. Yani aslında publisher veritabanından datanın snapshot’ının kopyalanmasıdır. Tüm değişiklikler tamamlandığında publisher üzerindeki değişiklikler subscriber’a real time olarak yollanır. LR ile sadece normal tablolar replike edilebilir. Materialized view, view, partition’un root tablosu yada foreign tabloları replike edilemediği göz önünde tutulmalıdır. Ayrıca şema bazlı veri aktarımı yapılamamakta. Farklı PostgreSQL versiyonları arasında LR yapılabilir. Birden fazla veritabanı üzerinden tek bir veritabanına LR yapılabilir. Farklı kullanıcı gruplarına replike olmuş datalar için yetki verilebilir. Veri objesi üzerinde yapılan her tekil değişiklik karşı tarafa aktarılır. Tek bir tablonun replikasyonu birden farklı cluster’lar üzerinde olabilir. LR i...

Hi, When PostgreSQL server and cluster versions don't match, PostgreSQL doesn't let you to run some functions, therefore you can not use efficiently and can not be advanced on it. To solve this problem you need to upgrade version or follow bellow steps. You can try all of steps for your development or test environments. I installed PostgreSQL server and client with different versions to show what is gonna happen If I try to dump. For sure, I can not get pg_dump! -bash-4.2$ pg_dump -s testdb > testdb_dump.sql pg_dump: server version: 9.6.3; pg_dump version: 9.2.18 pg_dump: aborting because of server version mismatch -bash-4.2$ pg_ctl stop -m fast waiting for server to shut down.... done server stopped -bash-4.2$ mv /var/lib/pgsql /tmp/pgsql_cluster_old mv: cannot move ‘/var/lib/pgsql’ to ‘/tmp/pgsql_cluster_old’: Permission denied To move pgsql file, need to be root at this moment. -bash-4.2$ logout [root@gunce-test tmp]# mv /var/lib/pgsql /tmp/pgsql_clu...

Hi, SQL Server announced Dynamic Data Masking feature not long ago and I'm Postgresql DBA! PostgreSQL does not have like that feature yet. So, what if we need data masking on our database's tables? If you need make strict your database tables and hide your column but not that all follow bellow function! This function shows only the first and last two characters. If you looking for some solution for your table's column and if it is the character you can create bellow function in your database and enjoy it! CREATE OR REPLACE FUNCTION dba.pg_datamask_column(col character varying)   RETURNS character varying AS $BODY$ begin  if (select length(col))<=2 then return col; elsif (select length(col))=3 or  (select length(col))=4  then return  (select replace(col,right(col,2),'xx')); else return (select  replace(substring(col from 1 for length(col)-2), right(substring(col from 1 for length(col)-2),length(col)-4),'xxxx')||right(col,2)); end if;...

Hello, I'm gonna share my experience on partitioning and show how affect partition on scripts. I'm gonna show you how to partitioning on existing table. Let's begin to following steps; Create new table. It's gonna be partitioned table. Create partition tables. Grant privileges to partition tables and partitioned table. Generate function and trigger on partitioned table. Insert data from main table to partitioned table. Create indexes if you had on main table on partition tables. Check partition tables everything is fine. Rename main table. Rename partitioned table name like main table Drop old main table. Not that: if you have some indexes or constraint on your main table, not to increase inserting time you should create your new main table and other partition tables without constraints and indexes. After insert whole data to partition tables you can add indexes and constraints. It will be help you to executing insert statement more quickly.  Bef...

Hi, If you can not reach pg_ctl command in bash also you are sure this command exists in your server, check $PATH. -bash-4.2$ echo $PATH /usr/local/bin:/bin:/usr/bin:/usr/local/sbin:/usr/sbin So, the directory of pg_ctl is not seem in PATH. Input directory of pg_ctl and check it then run it.

Hello, I would like to share my knowledge about how to get basebackup in PostgreSQL and with getting lots of errors and how to overcome it. So you should be ready to getting crash and doing stupid things and to be achieve! First of all, you need to make sure if your WAL archiving is enabled. You can look at Postgres.conf under cluster data file. Following steps you can connect your database as super user to continue but I'll use psql in bash. Access your environment and connect as postgres user and continue as bellow; [root@gunce]#sudo su - postgres -bash-4.2$ psql -d backup_test -c "select pg_start_backup(' backuptest ');"  pg_start_backup -----------------  0/F000028 (1 row) -bash-4.2$ tar -czf gunce.tar.gz data/* -C /var/lib/pgsql/9.6/ -bash-4.2$ psql -d  backup_test  -c "select pg_stop_backup();" NOTICE:  pg_stop_backup complete, all required WAL segments have been archived  pg_stop_backup ----------------  0/F000130 (...