On 11/7/22 09:43, Tom Lane wrote:
> Ron <ronljohnsonjr@gmail.com> writes:
>> On 11/7/22 08:02, Tom Lane wrote:
>>> call. It'd still be recommendable to pg_dumpall and restore into
>>> a freshly-initdb'd cluster, because otherwise you can't be real
>>> sure that you identified and cleared all the data corruption.
>> Why *just* pg_dumpall instead of "pg_dumpall --globals-only" and parallel
>> (for speed) dumps of databases?
> [ shrug... ] Sure, if you like to make your life complicated.
I don't remember if OP specified the size of his cluster, but Size Matters.
Dumping a multi-TB database to an SQL text file, and then restoring it will
be a whole lot slower than doing multi-threaded dump/restore.
--
Angular momentum makes the world go 'round.