Parallel fscks on large filesystems ... wondering about maxdsiz setting...

krad kraduk at
Thu Oct 6 09:21:59 UTC 2011

On 5 October 2011 19:25, Jason Usher <jusher71 at> wrote:

> Old 6.4-RELEASE system.
> Two filesystems exist, each of which is on its own raid controller.
> (Background fsck is not workable for various reasons that are tl;dr.)
> So, theoretically, doing both fscks at the same time is workable, since
> each of them are on their own controller, and no disk/controller resources
> are shared.
> HOWEVER, due to the large size and dense inode usage, we are forced to set:
> kern.maxdsiz="4096000000"
> And my question is:
> If I run two fscks at the same time, do I need to up this to 8192000000,
> or is this a per-process limit and I can run several processes that big,
> while leaving the value at 4096000000 ?
> (16 GB of ram, so either way we're well below)
> Thanks!
> _______________________________________________
> freebsd-questions at mailing list
> To unsubscribe, send any mail to "
> freebsd-questions-unsubscribe at"

If this becomes a major issue for you why not upgrade to 9 when its out then
you can have softupdates with journaling and remove most cases where you
need to run fsck, or make the jump to zfs. zfs will obviously require a bit
more thought.

More information about the freebsd-questions mailing list