I switched to subversion from cvs for version control about 6 months ago. On the whole I’m very happy with subversion except for one thing…
If the machine hosting subversion (which is actually also my development laptop) kernel panics or loses power or is force rebooted my subversion repository will get corrupted. So badly corrupted in fact that ‘admin recover’ will completely fail to repair it. I’ve been keeping regular backups (using ‘admin dump’) to keep snap shots of my repositories but it turns out that incremental dumps are not very reliable. About a week or two ago I lost my entire repository including backups.
Losing a repository isn’t the end of the world for me - I’m just a one man band and don’t need to track changes between multiple developers. I can quickly re-import my working directory into a new subversion repository. It is a pain in the neck and one I never had when I used cvs. I also don’t understand how the corruption happens - if it happened during a operation I’d understand it more - and its not like I’m running a subversion server - I’m just directly accessing my repository using the ‘file’ URL.
So now I’m moving my subversion repository off of my development machine and onto my Mac OS X server (which has a battery backup). I’m now willing to put up with the slowness in access times (the server is hosting behind a cable modem) to trade off the inconvenience of repository corruption.
Add New Comment
Viewing 1 Comment
Thanks. Your comment is awaiting approval by a moderator.
Do you already have an account? Log in and claim this comment.
Do you already have an account? Log in and claim this comment.
Add New Comment