[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

RE: Complexity of svn update?

From: Miller, Eric <Eric.Miller_at_amd.com>
Date: 2007-10-31 20:40:00 CET

> For example, suppose I have a repository with 100.000 files. Will it
> take less time if all files are in the root directory (so the
> repository does not have any subdirectories) than if I have lots of
> directories and subdirectories where the files are scattered? If the
> time to process each .svn directory is constant, processing one large
> repository with everything in the root will take the same time as
> processing an empty repository, which is much less time than the large
> repository with the files scattered in lots of directories.

So you are really just wondering about the overhead of updating a
hierarchical directory structure vs a flat one?

Someone may correct me if I'm wrong, but reading/writing each
.svn/entries file and creating/removing locks throughout the hierarchy
would create additional disk thrash. AFAIK the optimal setup would be a
flat structure.

Should be easy enough to verify though - just create 2 repositories, one
with 100k files in a directory and one with 100k directories with 1 file
each and compare update times.


To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Oct 31 20:40:47 2007

This is an archived mail posted to the Subversion Users mailing list.