How to repair an RPM database

Introduction

Incomplete or inconsistent data may be written to the rpmdb, causing corruption when a crash event occurs for an RPM or YUM transaction that was updating the RPM database. The reason for the crash could be a variety of reasons such as killing the process prematurely, resource or disk issues. 

A crashed RPM database can cause a myriad of issues related to package-management. Software such as EasyApache4 and YUM will also not work until the RPM database is in a functional state.

Due to this, one of cPanel & WHM's tasks during update sessions includes performing a check of the database and attempting a repair if needed. However, sometimes you may need to run this process manually.

Procedure

There are two ways to perform this check, one via the interface and via the command-line.

• Interface:
  • Navigate to the following tool via the WHM and click proceed to start the process
    • WHM >> Rebuild RPM Database

• Command-line:
  • From the command line, while logged in as root, you may simply execute the following script
    • /scripts/find_and_fix_rpm_issues

    • USAGE
       scripts/find_and_fix_rpm_issues [--findonly] [--rebuildonly] [--help]
      
      DESCRIPTION
       Detects problems with the rpm database and will rebuild the database if it
       detects problems. Exits 0 if RPM is working properly, or if we were able
       to fix it by rebuilding its database.
      
      This script also detects duplicate cPanel RPMs, removes, and then
       re-installs them if necessary.
      
      --findonly - Detect and report problems. Do not make any changes.
      
      --rebuildonly - Unconditionally rebuild the RPM database.

       

If you ever experience that the script times out or does not respond, you may want to view this article which describes possible causes.