Re: Large number of object files

From: Jeff King <peff@peff.net>
Date: Oct 24, 2012, 07:11 UTC
Message-ID: <20121024071109.GB29403@sigill.intra.peff.net>
In-Reply-To: <CACsJy8CcMBJLV=urVoWOQABQzQkC6y35spPF+_3fW6dtJjHtvg@mail.gmail.com>

On Wed, Oct 24, 2012 at 01:59:16PM +0700, Nguyen Thai Ngoc Duy wrote:

Show 8 quoted lines

> On Wed, Oct 24, 2012 at 12:21 PM, Uri Moszkowicz <uri@4refs.com> wrote:
> > Continuing to work on improving clone times, using "git gc
> > --aggressive" has resulted in a large number of tags combining into a
> > single file but now I have a large number of files in the objects
> > directory - 131k for a ~2.7GB repository.
> 
> Can you paste "git count-objects -v"? I'm curious why gc keeps so many
> loose objects around.

Presumably ejected from the pack because they are now unreachable. That's a rather large number, but if there was recent ref maintenance (e.g., deleting branches or tags), it is not impossible.

Show 5 quoted lines

> > Any way to reduce the number of these files to speed up clones?
> 
> An easy way to get rid of them is to clone the non-local way.
> Everything will be sent over a pack, the result would be a single pack
> in new repo. Try "git clone file:///path/to/source/repo new-repo".

If you have git v1.7.12 or greater, you can also use the "--no-local" option to clone. But as you mentioned, pruning is probably the most sensible thing (and for a non-local clone, those objects should not impact performance at all, as we will never even look at unreferenced objects).

-Peff