Re: migration of CVS repositery from one machine to another

[Jim Hyslop]

Todd Denniston wrote:
[...]
> moving a repo is covered in the manual[1].
> basically:
> 0) get everyone to do a final checkin and then release their sandboxes
> (mainly prevent confusion later).
> 1) Back up the repository [2].
> 2) take the backup to the other machine.
> 3) restore from backup.
> 4) configure the machine to serve cvs.

4a) test the configuration, including all scripts.

> 5) update everyone's CVSROOT values.

There is, of course, the possibility that someone might forget that the 
migration is happening, and check in a change to the old repository 
while the migration is in process. To eliminate this possibility, just 
create an empty file named CVSROOT/writers between steps 0 and 1. This 
also adds step 6: delete CVSROOT/writers file.

Step 5 is actually a rather difficult task. In order to make future 
migrations easier, I recommend that your server's actual machine name is 
*not* the one that is published to the users. Instead, set up an alias, 
which your system's DNS resolves to the actual server. That way, you can 
eliminate step (5) in all future migrations.

IOW, set up an alias in your DNS entries, called cvs.yourcompany.com. A 
DNS lookup resolves cvs.yourcompany.com to the actual machine. Then, in 
future migrations, you simply perform all the steps above, except step 5 
now becomes: Change DNS entries to point to the new machine.

--
Jim