Um...
separate revision range at a time... -r 6000:HEAD ?
only individual directories at a time... path/to/dir -r 6000:HEAD ?
Or...
This is a bit out there but you could spin up and install subversion on a
4XL amazon EC2 instance, import it on this very, very fast (virtual)
machine, then the moment it is done, tar up the repository directory,
download it and use that... with that said that will probably be fairly
expensive to do but it will save you a lot of time. If money is no object
get a maxed out quad G5 with 64GB of memory. The ASF will fit in ram.
There's a financial solution to every problem.
On Thu, Jun 16, 2011 at 5:08 PM, Bruno Antunes <bema_at_dei.uc.pt> wrote:
>
> On Jun 17, 2011, at 24:59 , Geoff Hoffman wrote:
>
>
>
> On Thu, Jun 16, 2011 at 4:05 PM, Bruno Antunes <bema_at_dei.uc.pt> wrote:
>
>> Hi,
>>
>> As part of the work of my PhD thesis I need to load the ASF Subversion
>> repository into my own local repository in order to mine and extract
>> information from the repository without overloading the ASF servers.
>>
>> I have downloaded the repository dump and started loading it into my own
>> repository. But the repository is huge (~45GB), and loading it using
>> 'svnadmin load' will take me days (~15).
>>
>> I tried 'svndumpfilter' to filter out some projects but I get the error
>> 'svndumpfilter: Unsupported dumpfile version: 3'. I'm using 'svndumpfilter'
>> version 1.6.12. Is there any way to overcome this error?
>>
>> Do you know any faster way to load the dump file or to filter out some
>> projects/revisions so I can speed up the process?
>>
>> Thank you in advance.
>>
>> Best regards,
>> Bruno Antunes
>
>
>
>
> Just a thought... Do you need the revision history or only the current
> (head) revision?
>
> Guessing if you do not need the revision history then it will be much
> smaller and faster to svn export their-stuff -r HEAD
>
>
> I need the entire revision history, because I need to extract historical
> information from the repository.
> But I could use the revision history from specific projects only, I don't
> need the entire repository, which contains all the ASF projects. The problem
> is I can filter these projects in the dump file.
>
> Best regards,
> Bruno Antunes
>
Received on 2011-06-17 02:17:32 CEST