2 of 2 people found this helpful
We always run with --no-config. With --no-config, it just restores the data which consists of both the Postgres metadata (i.e. workbooks, datasources, users), and the extracts (.tde pre-10.5, .hyper 10.5+). When you run without this --no-config I think it messes with the config such as IP's, special settings, etc. which is typically what we don't want because our non-prod environments are already configured the way we want them, and we just need to restore the data.
Many thanks. One of my colleagues identified what the problem was, and it wasn't this switch, so all good now.
Hey Chris. Can you share what the problem was, so that we can perhaps learn from it?
Basically, we maintain our DR server using a daily backup and restore process, with the restore having the --no-config switch.
It appeared that some of what I would have classed as config was getting updated by the restore.
What actually had happened was the restore was failing and only partially refreshing the data, and in an attempt to fix the issue, a restore was done without the switch, which then overwrote the config elements.
We have now figured out how it all happened, and corrected the issue that was preventing the restore working as expected, so all is now back up and running!