I noticed a little while ago somebody posted to say kjournald was
taking up a large amount of cpu, and I saw similar behaviour last
night. When I first set up journalling, I journalled two filesystems
(/usr/local and /home) without any problems. When I then attempted to
journal /usr, it kept refusing to unmount /usr (saying it was busy).
lsof took ages, (and I had to kill it ), so I did a top to find out
what processes where using /usr. Although I was in single user mode,
I noticed that the load on the box was 2 point something, with two
kjournald processes taking up 49% of cpu each. I couldn't see
anything using /usr, so I rebooted "init=/bin/sh", and was able to
journal /usr (and my / partition too) without further problems.
I made the journal files for the first two filesystems (and remounted
them as ext3) in runlevel 3, but brought the box single user to do
/usr, which was when I noticed the problem. I don't know if that's
relevant, and I haven't seen that behaviour since, so I can't be more
specific, I'm afraid.
I have two ide hard drives (with dma enabled), and apart from that, my
laptop is fairly standard (PII, cdrom etc, all ide). If you
need/desire further info, please tell me what you want.
I have done a few test power cycles, and the recovery seems to be
working fine. I would like to be able to test this on my alpha, but
none of the kernels that you supply patches for are particularly good
on alpha, and the patches don't seem to work against 2.2.13 (which
is). Are there any plans to port the patches (or should I have a go
myself)?
Thanks for your hard work (and showing me it all working on your
laptop at linux expo). I hope the above is helpful in some way.
Sean
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/