On Mon, 29 Jan 2001, Chris Colomb wrote:
> Exactly. Which is why we use SGI, AIX, and Solaris for that. Because
> they do journaling right and there isn't a performance impact.
Of course there's a performance impact. It may not be significant compared
to processor/disk throughput, but there *is* one. Keep in mind that the
systems you are talking about have been *hardware* optimized for
enterprise-level tasks, whereas most Linux systems are commodity Intel
boxes.
That said, there is definitely a case to be made for journaling on large
filesystems, despite the performance impact (which *may* be pretty
minimal, depending on what kind of usage your putting your filesystem to).
Any system that really requires guaranteed integrity and minimal reboot
time (e.g. a large, mission-critical database) is a good candidate for
journaling.
There are several journaling systems available for Linux right now.
Reiserfs has been available for some time, while journaling systems from
SGI and IBM are available, at least at the development level (although I
have no idea if they're ready for prime time yet), and there's work being
done on ext3.
My guess is that if you need enterprise-level features, you probably need
enterprise-level hardware, too, so this isn't really just about
journaling. Remember, the OS is only half of the server equation. :)
--
Todd A. Jacobs
Senior Network Consultant
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
