[tahoe-dev] Running multiple 'tahoe backup'?

Brian Warner warner-tahoe at allmydata.com
Sun Feb 22 22:36:53 UTC 2009


On Sat, 21 Feb 2009 18:47:42 +1100
Andrej Falout <andrej at falout.org> wrote:

> when I run single 'tahoe backup' command, my line utilization
> upstream is about 5%.
> 
> I am wondering if it is possible/acceptable to split data into several
> locations, and then run 'tahoe backup' of each in parallel?
> 
> I'm hoping it would allow me to better utilize my connection, and
> complete backup faster.
> 
> Is there any reason why this would not work?

Nope, no reason. That approach should be a good+easy way to increase the link
utilitzation. The only drawback would be the complication of managing
multiple independent Backups/ directories instead of a single one.

I'd advise keeping the separate processes from working on overlapping
directories, so that you don't get multiple 'tahoe backup' commands
working on the same file at the same time. I'm not yet sure that the
backupdb is robust against parallel manipulation of shared pathnames.
(I think it should be robust against parallel manipulation of disjoint
pathnames, though).

cheers,
 -Brian



More information about the tahoe-dev mailing list