You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Bill Moseley <mo...@hank.org> on 2006/09/29 19:34:57 UTC
Recommendation for creating a repo of "docroot"
I asked this on #svn so sorry if you already saw this. This is kind
of a general "what would you do?" question.
I'd like to move a client's website to subversion. They have all been
using FTP to update the site, and for a number of obvious reasons I
think they would be better off keeping working copies on their own
machines/accounts than updating the docroot directly as they are now.
My concern is that the content is now about 5GB and about 6000 files
and a mix of mostly binary and some text content, and I'm sure most of
them don't want to check all that out onto their laptops.
So, I'm looking for suggestions how to create the repository. Should
the repository contain everything, or just the text files that they
will be mostly interested in updating?
If just a sub-set of files, then how best to create that repo?
Manually copy the 1700 or so text files to a new directory tree, then
import that tree? Then cp the old docroot on top of the new tree and
go crazy with svn ignore? That could be a bit of work just keeping
the svn ignore properties updated.
What would you do in this situation?
Thanks,
--
Bill Moseley
moseley@hank.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Recommendation for creating a repo of "docroot"
Posted by James <ho...@outofcontrol.ca>.
On Sep 29, 2006, at 3:56 PM, Bill Moseley wrote:
> On Fri, Sep 29, 2006 at 03:42:32PM -0400, James wrote:
>> On Sep 29, 2006, at 3:34 PM, Bill Moseley wrote:
>> We currently do all our sites using this method. Our largest site
>> currently is 41,770 files for a total of 250MB. We develope locally
>> on our Macs and check in changes as we need, using mod_dav_svn.
>> Albeit a bit slow when updating the entire try (roughly 1 minute) we
>> find this works quite well for us. Haven't tried this on anything
>> like 5GB yet.
>>
>> Our file mix is mostly small images and a lot of php scripts.
>
> So everything (or most) is in your repo? My client has these hundred
> megabyte webcasts (quicktime movies -- not a great thing to svn co.
>
> Do you make use of svn:ignore on the clients? Or does everyone have a
> full copy of the repository?
>
Our site is updated on a daily basis from a web-interface by our end
clients. Our developers work on the site locally on their macs, then
update the repos, which in turn updates the dev server automatically
via a post-commit, once every day or so. In order for the dev site
and the local sites to properly be tested we needed the images and
end client files to be present.
All that to say that we tried using svn:ignore on the end clients
files, but ran into problems properly testing our code before going
live. So no, we do not use svn:ignore on any files and everything is
in the repos, with the exception being our MySQL DB's.
We realize this is not the best setup, as having all those files and
images in the repo is a bit uncouth, but it does work well none the
less. We do not use BerkeleyDB and use mod_dav_svn instead of svn://.
We hope to try out our setup with the BerkeleyDB to get a speed
improvement on commits.
James
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Recommendation for creating a repo of "docroot"
Posted by Bill Moseley <mo...@hank.org>.
On Fri, Sep 29, 2006 at 03:42:32PM -0400, James wrote:
> On Sep 29, 2006, at 3:34 PM, Bill Moseley wrote:
> We currently do all our sites using this method. Our largest site
> currently is 41,770 files for a total of 250MB. We develope locally
> on our Macs and check in changes as we need, using mod_dav_svn.
> Albeit a bit slow when updating the entire try (roughly 1 minute) we
> find this works quite well for us. Haven't tried this on anything
> like 5GB yet.
>
> Our file mix is mostly small images and a lot of php scripts.
So everything (or most) is in your repo? My client has these hundred
megabyte webcasts (quicktime movies -- not a great thing to svn co.
Do you make use of svn:ignore on the clients? Or does everyone have a
full copy of the repository?
Thanks,
--
Bill Moseley
moseley@hank.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Recommendation for creating a repo of "docroot"
Posted by James <ho...@outofcontrol.ca>.
On Sep 29, 2006, at 3:34 PM, Bill Moseley wrote:
> I asked this on #svn so sorry if you already saw this. This is kind
> of a general "what would you do?" question.
>
> I'd like to move a client's website to subversion. They have all been
> using FTP to update the site, and for a number of obvious reasons I
> think they would be better off keeping working copies on their own
> machines/accounts than updating the docroot directly as they are now.
>
> My concern is that the content is now about 5GB and about 6000 files
> and a mix of mostly binary and some text content, and I'm sure most of
> them don't want to check all that out onto their laptops.
>
> So, I'm looking for suggestions how to create the repository. Should
> the repository contain everything, or just the text files that they
> will be mostly interested in updating?
>
> If just a sub-set of files, then how best to create that repo?
> Manually copy the 1700 or so text files to a new directory tree, then
> import that tree? Then cp the old docroot on top of the new tree and
> go crazy with svn ignore? That could be a bit of work just keeping
> the svn ignore properties updated.
>
> What would you do in this situation?
We currently do all our sites using this method. Our largest site
currently is 41,770 files for a total of 250MB. We develope locally
on our Macs and check in changes as we need, using mod_dav_svn.
Albeit a bit slow when updating the entire try (roughly 1 minute) we
find this works quite well for us. Haven't tried this on anything
like 5GB yet.
Our file mix is mostly small images and a lot of php scripts.
James
>
> Thanks,
>
>
> --
> Bill Moseley
> moseley@hank.org
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: users-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org