tlug: Journaling file systems question (Re: Journaling Filesystems)

[Fredric Fredricson <fredric.fredriksson@example.com>]

This debate puzzles me a little and since somebody must ask the
stupid questions, why not me ;-)..

I know very little about JFS but my impression was that they 
operated something like some database engines.

This is how I figured a JFS should operate (roughly):
1. A write to file creates a "transaction record"
   (with database terminology) on disk.
2. The "transaction record" is read and a modified
   copy of the file is created.
3. The the file is replaced with it's copy.
4. The "transaction record" is removed.

A read always checks the "transaction records" first
to make sure the read is consistent with earlier
writes.

I realize that this is not exactly "journaling" but
more like a "safe" file system. It also fits very bad
with the "data" and "metadata" stuff.

Can somebody explain how wrong I am here?

/Fredric (just curious!)
--------------------------------------------------------------------
Next Technical Meeting: March 11 (Sat) 13:30 Temple University Japan
* Topic: "What's new in Perl 5.6"
Guest speaker: Simon Cozens (TLUG Perl guru)
Next Nomikai Meeting: April 20 (Thu) Linux Conference 2000 Spring Ed.
--------------------------------------------------------------------
more info: http://www.tlug.gr.jp        Sponsor: Global Online Japan