CentOS系统PostgreSQL创建数据库集群执行initdb报错：cannot be run as root

DevOps 小满

2026-5-19

Fixing "initdb: cannot be run as root" on CentOS

Hey there, I’ve dealt with this exact error plenty of times when setting up PostgreSQL on CentOS—no worries, it’s a straightforward fix rooted in PostgreSQL’s security best practices (it intentionally blocks root to prevent running the database with excessive privileges). Here’s how to get past it:

Step 1: Create the PostgreSQL system user (if it doesn’t exist)

Most CentOS PostgreSQL packages create the postgres user automatically, but if you’re building from source or it’s missing, run this as root:

useradd postgres
# Optional: Set a password for the user (not always required for local operations)
passwd postgres

Step 2: Switch to the `postgres` user

You need to run initdb as the dedicated PostgreSQL user. Use this command to switch to their shell (the - ensures their environment variables are loaded):

su - postgres

Step 3: Run `initdb` with your target data directory

Now execute initdb and specify where you want to store the cluster data. The default path for CentOS is /var/lib/pgsql/data, but you can use any path you prefer (just make sure the postgres user owns it first).

For the default path:

initdb -D /var/lib/pgsql/data

If you’re using a custom directory, first create it and set permissions (run this as root before switching users):

mkdir -p /path/to/your/custom/data/dir
chown -R postgres:postgres /path/to/your/custom/data/dir

Alternative: Run directly without switching shells

If you don’t want to switch to the postgres shell, use runuser to execute the command as them (great for scripts):

runuser -l postgres -c 'initdb -D /var/lib/pgsql/data'

Pro Tip for YUM-installed PostgreSQL

If you installed PostgreSQL via CentOS’s official repos or PostgreSQL’s YUM repo, you can skip the manual initdb and use the setup script, which handles all the user permissions automatically: