regressing destination - out of temp space

[Ed Greenberg]

Hi Folks,

I have rdiff-backup  1.2.8-13.el7  from epel repo on centos 7.  I've been
using it for years, and it saves my butt regularly.

I have a failed backup. The backup directory is 247 gb.  The root, where
/tmp lives is 8GB with 6 free.   When rdiff-backup tries to regress the
destination, it runs out of space, and it's the tmp directory that is
running out of storage.

Two questions:

1. How much /tmp do I need to regress the 246gb directory?
2. Would it be better to run --check-destination-dir on the backup server
or let the rdiff-backup on the client do it's job?

Thanks,

Ed Greenberg