On Mon, Nov 15, 2010 at 14:50, Seak, T. F. <lapsap7+svn_at_gmail.com> wrote:
> On Mon, Nov 15, 2010 at 19:44, Andy Levy <andy.levy_at_gmail.com> wrote:
>>
>> You cannot draw any conclusions about the size of an export or
>> checkout from the size of the repository database itself.
>
> Why not? Those values I wrote are NOT repository database size, but
> ... how should I say?.... the size of the folder after EXPORT.
You stated: "One repository is 183 MBytes but transferred is only 111.89 MBytes
Are you counting the size of the repository on the server, or the size
of the resulting export?
> On the other hand, the written "transferred" shown in TSVN should mean
> something. The question is to know what that means exactly.
>
>> I'm fairly sure that the HTTP layer uses compression. What will
>> disabling it gain you?
>
> I'm doing performance test because our SVN server is serving a lot of
> off-site workers (in other countries). They have been asking us (who
> are working in-site) to enable compression. We've answered that it's
> already enabled but they just keep on asking. If I could disable, I
> *gain* confirmation and confidence from off-site workers that we've
> already done all's possible for them to make their lives easier.
Compression is configurable by the client - if they want to know for
certain that compression is being requested, they should be setting
http-compression = yes in %appdata%\subversion\servers
But how often are they doing this in the first place? After the
initial checkout, they should only be performing updates, which will
only transfer diffs over the wire.
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=4061&dsMessageId=2682024
To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2010-11-15 22:07:56 CET