huge db/strings dissapearing after dump/load : Was: Clean out large file

After exporting every single revision of this repository I found out that ther
was no big file in this repository at all. (although the developer told be
that he commited a really big file)

I found out that every single revision wasn't any bigger than 4-5Mb, but the
db/strings file was about 600megs big.

I deleted all unused logs and ran 'svnadmin recover' without any success, the
repository kept wasting disk space.

Now what I finally did was exporting the repository via 'svnadmin dump -r
0:HEAD' and loading it back to an empty one via 'svnadmin load'

The results:
The etl_ci-89 is the huge original repository, while the etl_ci-load is the
newly created/loaded repository

subversion@turtle ~/etl_ci-save $ ls -lh
total 12M
drwxr-xr-x 8 subversion subversion 4.0K Jan 18 19:40 etl_ci-89
-rw-r--r-- 1 subversion subversion 12M Jan 18 20:15 etl_ci.dump
drwxr-xr-x 7 subversion subversion 4.0K Jan 18 20:17 etl_ci-load

subversion@turtle ~/etl_ci-save $ du . -h --max-depth=1
609M ./etl_ci-89
7.9M ./etl_ci-load
628M

Well thats about factor 100 smaller than before ;-)

So my question is how can this during normal operation happen?
I also checked and removed any dead transactions (there was one). Could a dead
transaction cause this?

Cheers,
Reinhard

On Tuesday 18 January 2005 18:43, Reinhard Brandstädter wrote:
> Hi,
>
> I've got a repository to which a user (dumb one) committed a really large
> file (serveral 100Mbs) by accident causing the repository to be quite big.
> Now I've got the problem that I don't know the exact file that was that
> big, nor do I know in which revision he commited the file?
> Is there an easy way to find the revision and the file (by size) and then
> clean out that file?
>
> Cheers,
> Reinhard

-- 
Reinhard Brandstaedter
Wiener Str. 28                  phone: 0043 699 12419541
4020 Linz                       icq:   73059068
Austria                         email: r.brandstaedter@gmx.at