The first section of this document describes how a persistent copy of the Vertica catalog is managed and laid out on disk. This information is useful for advanced Vertica users who are troubleshooting catalog disk space utilization issues.
The second section of this document describes how to run and interpret the output of the dump_allocator_usage API. This information is helpful for database administrators who are troubleshooting the reasons behind the growth of the in-memory catalog.
Keep reading it here:
https://my.vertica.com/kb/Troubleshooti ... atalog.htm