RE: Repository synchronization between two cvs server

[Arthur Barrett]

Vhari,

> can some one guide me on how to setup a ssh for CVS access, 
> we have 2 dev
> teams in 2diff locations
> n two teams would required to commit their changes. Existing 
> bandwidth is
> creating some issues while checking out code. 
> 
> looks like ssh can solve my proble.....

I think you have misunderstood the thread - ssh does not affect the
bandwidth required.

As I wrote previously:

> CVS was designed for use on networks much slower than 
> todays, and so therefore operates well in a multi site 
> environment without the need for repository replication.
> 
.
.
.
> However let me stress - CVS does not require repository 
> replication in order to allow teams of people at multiple 
> locations to efficiently use the server.  


Write to the group and tell us exactly what you mean by the phrase
"existing bandwidth is creating some issues while checking out code".
Specifically:
1. what is the command to check out code
2. what is the size of the resulting sandbox
3. what files are you checking out (largest file size, average file
size, number of files, type of files)
4. what is your server, what is your client (cvs versions and also os
versions)
5. how often do you run the "checkout" command?
6. how long does it take (elapsed time) to run the "checkout" command

I've previously seen issues where users complain of "slow checkout" over
WAN links only to find they have a 4GB sandbox, and 50 developers who
delete it on Friday night and all expect to "checkout" at 8am  on Monday
morning.  This is a process issue not an issue for the tool (ie: if you
modify what you do slightly, like use "update" instead of "checkout" or
use network compression then there is no need for tool/technology
changes).

Regards,


Arthur Barrett