This will address a few questions, as well as a possible failure to restart the database once shut down.
The cause for the possible database restart failure with SSH errors between nodes is the dradmin OS users password has expired. Testing ssh between nodes using the OS dradmin user presents a password expiration notice requesting the password be changed as it's expired.
All supported Performance Management releases
Who is the 'dradmin' user?
It is the OS level user that acts as the Data Repository Vertica database administrative user. It is often referred to as the 'dradmin' in user guides. This is the recommended user name per the installation guides but it can be set to a different name.
How is the user created?
The user can be created in one of two ways.
Is there a difference between the manually created user, and the user created by the installation process? The one difference is that the user generated by the installation process does not set a password for the user.
When allowing the installation to generate the user, you specify the dradmin user name in the dr.properties file. This file is then passed to the installation at run time. When the installation runs, if the user doesn't already exist it is created with no password. When generating the user manually prior to installation, its rare that the user would be created without a password.
What is required to change the password?
To change the users password, follow the normal Linux user password change options. Most common would be to:
Note that the password entry won't be echo'd to the command prompt. Type with care in order to avoid setting an unknown password.
After setting a new password for the OS user, the database restart failures with SSH errors should be resolve. If not there may be some other problem with the SSH keys themselves.