[p4] p4 obliterate - Michael Vance - com.perforce.perforce-user

7 messages in com.perforce.perforce-user[p4] p4 obliterate

From	Sent On	Attachments
Michael Vance	12 Nov 2003 18:44
Matthew Rice	13 Nov 2003 05:36
Tomm...@ni.com	13 Nov 2003 07:16
Dave Lewis	13 Nov 2003 07:37
Ivey, William	13 Nov 2003 07:39
Grills, Jeff	13 Nov 2003 10:38
Michael Vance	13 Nov 2003 15:47

Subject:	[p4] p4 obliterate
From:	Michael Vance (van...@treyarch.com)
Date:	11/13/2003 03:47:28 PM
List:	com.perforce.perforce-user

This is the thing I've found, too. My guess is it's linear with respect to db.rev, but not your depot size. It takes me about 15 minutes to do an obliterate.

The other question I have is this: sometime ago I read about a trick involving replacing the foo.gz files in the file's revision history on the server with a text file analagous to 'this moved', but I can't seem to find reference anymore. Anyone recollect where this is?

Thanks,

Ivey, William wrote:

-----Original Message----- From: Michael Vance [mailto:vance at treyarch.com]

We have some dead data in branches that were never used, etc., totaling several GB, that I would like to obliterate. I've done a test of the obliterate command on my db on a couple (three) test files and I noticed that a) it locks the db, b) it takes really, really, really long. And this is just one some three test files, I want to delete 100s of thousands of files. Is this even feasible?

If it helps, I haven't found the obliterate time to be linear - that is, if it takes time t to do one file, it doesn't take 100t to do 100 files. I have obliterated several thousand files at a time. (I don't recall how long it took, but I think it was on the order of 5 minutes, maybe a little more - some of that was just the command listing what it had done.)

Of course, the operations that have been performed on the files prior to the obliterate have a lot to do with it, I'm sure. -Wm