8. Moving the AccuRev Server and
Repository to Another Machine

The AccuRev data repository should be physically located on the machine that runs the AccuRev Server process. (This is not an absolute restriction — see READ ME NOW: Assuring the Integrity of the AccuRev Repository on page 5.) The repository consists of multiple slices: the site slice contains information that pertains to the entire repository, and each depot has its own slice.

From time to time, you may want (or need) to have the AccuRev server process run on a different machine. To accomplish this, you must:

•	Install AccuRev on the new machine.

•	Move your data repository to the new machine.

The procedure described below is safe to use even if the new machine has a different byte order than the old machine.

Procedure for Moving the Repository

Make sure you perform each of the following steps on the appropriate server machine. We call them:

•	The source machine — where the AccuRev server is currently running and the data repository is currently located.

•	The destination machine — the machine to which you want to move the data repository.

Note: the steps below always show UNIX/Linux pathname separators ( / ). When you’re executing commands on a Windows machine (either source or destination), be sure to use Windows pathname separators ( \ ).

The procedure calls for multiple stops and starts of the AccuRev Server process. For details on how to accomplish this, see Controlling Server Operation on page 21.

On the Source Machine

1.	Execute the accurev show slices and accurev show depots commands, and save the output for reference in the following steps.

2.	Perform a backup of the AccuRev repository, as described in Backing Up the Repository on page 6.

3.	Make the backed-up files accessible to the destination machine.

On the Destination Machine

1.	Request a new license key for the destination machine by filling out the form at http://www.accurev.com/license-request.shtml. Save this key file to a location on the destination machine.

2.	Perform a full AccuRev Server installation on this machine as described in the AccuRev Installation and Release Notes. The installation directory pathname need not be the same as on the source machine.

3.	Stop the AccuRev Server, as described in Controlling Server Operation on page 21.

4.	Perform a restore of the AccuRev repository you backed up on the source machine, as described in Restoring the Repository on page 8.

Note: Do not copy over the /storage/db directory on the destination machine. This directory should not have been included in the backup of the source machine repository, and you do not want to overwrite the /storage/db directory on the destination machine.

Check that the file ownership and permissions are correct; they may need to be altered if the username used to run the AccuRev server is different on the source and destination machines, or if your backup/restore tool does not correctly preserve these settings.

5.	Update the server properties:

maintain server_properties update

See The ‘maintain’ Utility on page 115 for more information.

6.	Change the locations of the slices to their locations on this machine using the output from the accurev show slices and accurev show depots commands on the source machine. To do so, run the following command for each slice:

maintain chslice <slice-number> <new-location>

7.	Start the AccuRev Server.