Messages in this thread | | | Date | Sun, 01 Mar 1998 01:18:26 -0600 | From | Trever Adams <> | Subject | Re: Toronto Filsystem? |
| |
> From: "Benjamin C.R. LaHaise" > Date: Sat, 28 Feb 1998 19:14:09 -0500 (EST) > Subject: Re: Toronto Filsystem? > > On Sat, 28 Feb 1998, Albert D. Cahalan wrote: > .... > > I think this could get ugly when someone appends to a 1000 MB file. > > You might want an inode hack that indicates "blocks in this file > > before #345987 are specified by inode 42 on the layer below". > > Actually, this is easier than it seems. A simple restriction that makes > everything work well is to require that all underlying layers be made > read-only, which avoids a mess of consistency problems. Then, use a top > layer that supports sparse files. Reading a file becomes: try layer one - > oh, not there; try layer 2... No, I don't have working code yet. > > -ben > Not cool at all. Once the original file owner updates the file (say the group agrees on a new standard, or root updates the file) any way but clean implimentation and write of the entire modified file becomes corrupted. You cannot ignore the possibility of under layers being modified, especially if you want multiple layers (beyond 2). This is true with spares files or the way the way suggested by the post you responded to. Once an upper layer diverges (by modification not by "WHITE OUT") it must forever remain independent for that file. I see no way around it.
As always, if I am wrong, please correct.
~~~~~~~~~~~~~~~~~~~ Trever Adams arabian@onramp.net http://rampages.onramp.net/~arabian/silent.htm ~~~~~~~~~~~~~~~~~~~ "No man knows how bad he is till he has tried very hard to be good." ~C.S. Lewis
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |