Berkeley db corruptions when running svnserve and viewcvs (?)

[Stefan Urbat <st...@apastron.lb.shuttle.de>]

Yesterday I managed to make viewCVS run together with subversion. We use
svnserve as standonly server on the same machine, to handle checkout and
commit requests and the like solely, but yesterdays data corruptions make
me wonder, if it can be dangerous to run the view-only Apache/viewCVS (on
Apache 1.3) together with svnserve for the described environmen?

Details: Debian Woody GNU/Linux server, running repositories on xfs with
recent 2.4 Kernel, some clients don't work with http/webdav (while working
fine with svnserve), namely Solaris 9 and AIX 4.3.3, therefore we can't
use the more flexible http generally for checkout and so on too (yet).

Any hint/opinion is highly appreciated, we work since Monday with a dozen
or so developers on subversion instead of cvs after successful migration;
the performance issue with checkout and update seems to me more in the
dated server hardware...


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Berkeley db corruptions when running svnserve and viewcvs (?)

[Nathan P Sharp <ns...@phoenix-int.com>]

Most likely this, have your read it yet?

http://svnbook.red-bean.com/svnbook/ch06s05.html

Stefan Urbat wrote:

>Yesterday I managed to make viewCVS run together with subversion. We use
>svnserve as standonly server on the same machine, to handle checkout and
>commit requests and the like solely, but yesterdays data corruptions make
>me wonder, if it can be dangerous to run the view-only Apache/viewCVS (on
>Apache 1.3) together with svnserve for the described environmen?
>
>Details: Debian Woody GNU/Linux server, running repositories on xfs with
>recent 2.4 Kernel, some clients don't work with http/webdav (while working
>fine with svnserve), namely Solaris 9 and AIX 4.3.3, therefore we can't
>use the more flexible http generally for checkout and so on too (yet).
>
>Any hint/opinion is highly appreciated, we work since Monday with a dozen
>or so developers on subversion instead of cvs after successful migration;
>the performance issue with checkout and update seems to me more in the
>dated server hardware...
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
>For additional commands, e-mail: users-help@subversion.tigris.org
>
>  
>

-- 

 Nathan P Sharp
 Sr. Software Engineer, Phoenix Integration
 540-961-7215 x 309
 nsharp@phoenix-int.com


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org