Re: journaling filesystem

The Mitochondrion (
Thu, 14 May 1998 12:35:49 -0400 sez:
> It doesn't matter too much what it is, as long as it can be used to build a
> consistency/rollback mechanism with 100% reliability. Efficiency is good,
> hence the ideal of being able to impose a partial ordering on block writes,
> but anything would do. After all, journalling file systems are not generally
> chosen for their speed, are they?

Actually, from the early papers I read about Journalling filesystems,
their whole purpose was efficiency, with other possible features (like
audit trail) as good but ancillary features.

As I remember it, with good read-caching, you need very little disk
activity for reads, and since the journal is written sequentially to
the device, you don't have to do seeks very often on writes. One
author was claiming speedups of 2-3 orders of magnitude -- but that
was a few years ago. I can dig up the references if anyone is


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to