[mb-datafeed] IMPORTANT: Schema change update

[Robert Kaye]

Hi!

We updated the server today with a schema change update, which means 
that if are running a replicated MusicBrainz database, you will need to 
update your mb_server codebase in order to continue updating your live 
data-feed. NOTE: If you do not do this, your database will no longer be 
up to date!!!

Here are the instructions for the impatient on how to update your 
server:

<turn off your cron job that replicates packets>
cd /tmp
wget ftp://ftp.musicbrainz.org/pub/musicbrainz/data/puid.dat
wget ftp://ftp.musicbrainz.org/pub/musicbrainz/data/puidjoin.dat

cd <mb_server location>
cvs -z9 update -r RELEASE_20060310-BRANCH -dP
./upgrade
cd admin/replication
./LoadReplicationChanges
cd -
<turn on your cron job that replicates packets>

And now more detailed instructions:

0. Turn off the cron job that runs your replication. Usually this is 
done via:

     crontab -e
     < put a # before the line that runs LoadReplicationChanges >

1. Log in to your machine that hosts the replicated database.
2. Download these two files to your /tmp directory:

      ftp://ftp.musicbrainz.org/pub/musicbrainz/data/puid.dat
      ftp://ftp.musicbrainz.org/pub/musicbrainz/data/puidjoin.dat

3. cd to the top-level mb_server directory (the directory where the 
musicbrainz server code is located)
4. Download the latest cvs source code from the RELEASE_20060310-BRANCH 
branch by issuing this command:

      cvs -z9 update -r RELEASE_20060310-BRANCH -dP

5. Review the update of the source code. If you have made any changes 
to your local repository, you may need to inspect the download for 
merge conflicts. Resolve the conflicts as necessary before proceeding 
to the next step.
6. In the same directory, run the upgrade script:

      ./upgrade

This will load the data files downloaded to /tmp and import then into 
the database.

7. Run the latest batch of replication changes, to make sure all is 
well:

     cd admin/replication
     ./LoadReplicationChanges
     cd -

8. Turn on the cron job that runs your replication:

     crontab -e
     < remove the # before the line that runs LoadReplicationChanges >

That should be it. If you have any problems post your problems 
questions here. So far I've managed to upgrade three machines without a 
hitch. Good luck!

PS. Our scripts hiccuped a little during the upgrade and skipped one 
schema version number. We are now at version SEVEN.

--

--ruaok              Somewhere in Texas a village is missing its idiot.

Robert Kaye     --     rob at eorbit.net     --    http://mayhem-chaos.net