Subject: Re: `Large Inodes'
To: None <eeh@netbsd.org>
From: Bill Studenmund <wrstuden@nas.nasa.gov>
List: tech-kern
Date: 03/29/1999 10:26:44
On Fri, 26 Mar 1999, Eduardo E. Horvath wrote:

> On Fri, 26 Mar 1999, Bill Studenmund wrote:
> 
> > Our system is a little different - we never blow away the file, just its
> > data. So each file always has an inode, and has 0 length. That's why a
> > larger inode would have been cool. :-)
> 
> Have you considered truncating the file to one fragment and stashing the
> extra data there and setting one bit in the inode indicating the file has
> been migrated and the information to track it is in that fragment?

The big problem with this idea is that we need to keep our private data
around all the time on a file, even when it's fully resident. For
instance, the modification times are kept seperate so that an ls -l
doesn't reveal the behavior of the restoration, only what the user
expects. :-)

Take care,

Bill