Obliterate Difficult to Use

[jcoccia]

I'm wondering if there are any tips that I am missing. I've searched the forums and haven't found much. The obliterate is a tedious process. I wish there was a way to reference files and folders by ID numbers so we could look up the current locations of the files.

Anyway, I've got this one file that I'm trying to obliterate. (Actually I have a lot of files and folders, but tracking down outdated branch references to renamed and deleted structures that no longer exist is getting very old.) This one file has no branch references, but the obliterate command shows the usual error:

All branches of an item must be obliterated before the item itself can be obliterated. It is possible that some of the branched items have been deleted, but not obliterated. You must obliterate all branches before you can obliterate this item. This item has branches at:

And of course, there is no reference.

<sigh> How can I make this obliterate process faster and easier. I mean, it's pretty bad, so I figured that I was missing something simple. Thanks.

[Beth]

Obliterate is not recommended for most customers. If you have to obliterate, then it can help to obliterate right after you make a delete. The reference not being found means that the branch it was trying to find was deleted already and will show up in the obliterate tree somewhere.

Other things that can help is to run a history search for branches. You can filter by the Action Obliterate in the options window that comes up. From that you can see what was branched from what before you obliterate.

Another thing you might try is our obliterate power toy posted on this page: http://www.sourcegear.com/vault/powertoys.html.

[jcoccia]

Thank you for the response. I did try the obliterate powertool, but it doesn't seem to add much additional functionality over the website admin page.

If obliterate is not recommended, once a file/project has been deleted, what is the preferred method to clean up the database?

[Beth]

Is there a reason why you want the data gone? Is your database very large? Are there any performance issues?

Obliterate is the only way to permanently remove data from a repository, but removing data goes against the idea of tracking the entire life cycle of a project. If the database isn't large, then deleting data is enough for most users.

Please note that the Vault Export Import tool doesn't work around obliterates as it requires an intact history.

[jcoccia]

We were using Vault to store documentation and related files to programming projects. We now have a document managment system for those. We figured that after those items were deleted from Vault that it is a good maintenance practice to remove those files from the database. History for these items was unnecesary. Over the years, Vault was simply used as a convenience for more than maintaining source code.

The database is not experiencing slowdowns, it's only 5.7 GB. We figure that 50% of that are files no longer maintained in Vault or are not source code files.

[Beth]

Obliterate isn't really a normal maintenance thing for Vault. It's necessary for users who have very huge databases that run them out of disk space or for users who are hitting the SQL size limits for the version of SQL they are using. As long as the files you don't need are just deleted, Vault will continue to perform well.

If you would prefer to obliterate, then the suggestions I made before can help. If you want a little more help with that, you can send me an email and we can work on that offline. Just send an email to support at sourcegear.com (attn: Beth) with a link to this forum thread.

[jcoccia]

thank you Beth. If it's not a normal maintenance and we're not having difficulties then I'm inclined to not worry about it. I'll inform the rest of the team of your recommendations. Thank you Beth.

[Beth]

You're welcome.