On Thu, Mar 25, 1999 at 02:00:48PM -0500, Thor Lancelot Simon wrote:

> I don't understand what you mean by "stability".  FFS with soft updates is
> exactly like normal FFS in terms of failure recovery, it just reduces
> synchronous write overhead while the filesystem is running.

Maybe I'm confused more than I thought, then... I thought soft updates did
the same thing as a JFS but for individual writes, where writes are planned
and then executed, after which the FS is updated to show that the write was
successfully completed.

This isn't the case? <boggle>

Thanks in advance for setting me straight...

> LFS will at least let you boot your system faster, at the expense of
> predictably losing the last (checkpoint interval) worth of updates
> (whereas in FFS you may lose _some unpredictable fraction of them_), by
> rolling back to the last checkpoint in the log.

Hm. Okay. So, with update set to sync pretty frequently, not only do you
not lose much that way, but you've localized where you have to look for
corruption. I guess that's cool, although (as mentioned in the 4.4BSD book)
physical media defects cropping up would seem to throw the world into
confusion. But then, I guess a JFS doesn't really have any innate defense
against that anyway. <more boggling>

-- 
Mason Loring Bliss             ((  "In the drowsy dark cave of the mind dreams
mason@acheron.middleboro.ma.us  ))  build  their nest  with fragments  dropped
http://acheron.ne.mediaone.net ((   from day's caravan." - Rabindranath Tagore