[clug] Kernel without initramfs

Daniel Pittman daniel at rimspace.net
Thu Mar 26 06:08:10 GMT 2009

David Cottrill <cottrill.david at gmail.com> writes:

> The ext3 was just to make sure it was all working - I'm planning on
> using ext2 for production.

That seems reasonable enough.

> Because I'm planning on using this for a single application server
> with no chance of data recovery if there is a failure - journalling
> doesn't seem too useful for me. In my research it popped up that the
> journal is updated every minute by default - another nail in the
> coffin of ext3.

... um, not that I am aware of.  There is a default, but tunable,
journal commit every five seconds, which results in some journal
activity, but that is not entirely distinct.

Also, not going to do anything unless you actually dirty data, in which
case you are going to have to write it out some time anyhow.

If you have a lot of short-lived temporary files that are deleted after
use you can tune the journal commit interval to avoid capturing them in
the forced commit.

> I have considered XFS, but I'd have to look into the specifics of its
> journalling as well.

I don't believe it has anything like the feature you describe but, then,
I don't entirely understand what your concern is anyhow.

Um, if you really have no data worth preserving I would simply run it
out of a tmpfs, using a union filesystem to stack on top of the real
content, and ignore the rest of the comments.

If you really need more space than just RAM can provide then dedicate
some flash space to swap, which is not journalled, and allow tmpfs to
swap pages to that under memory pressure.

It may be, in fact, that you are already doing this, in which case the
choice of underlying filesystem is irrelevant: since nothing writes to
it you won't really distinguish between the choices, in general.

> Still can't get my non initrd kernel to boot but I've gotten
> sufficient speed just from better resarch into the boot process.

Good.  What was the cause of the delays in the end?  I don't presently
have any such need, but it is always interesting to know how to improve
boot speed with live systems.


More information about the linux mailing list