tech-kern archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: zfs questions
Greg Troxel <gdt%lexort.com@localhost> writes:
> Takashi YAMAMOTO <yamt9999%gmail.com@localhost> writes:
>
>> * is anyone using zfs in netbsd seriously? :-)
>
> I am using it on my desktop, and a on a dom0 hosting pkg builders.
>
> It mostly works, except that one must be careful not to simultaneously
>
> run low on memory
> run programs that dirty mmap'd pages (syncthing!)
>
> and probably also not
>
> run really low on memory such that there is a lot of paging
>
This is a reaonsably summary of the quirks.
I use ZFS all of the time on two of my build systems, one for pkgsrc and
one for the OS. Back when I had a $DAYJOB, I used ZFS on Solaris and on
Triton Data Center and find that most of the features I am familar with
work on NetBSD. Mostly just with the quirks mentioned above.
>
> As I undertand things, from chuq@ figuring out a deadlock after much
> offlist discussion with Brad, there is a deadlock from trying to free
> pages from the arc and trying to write dirty pages. The fix is to have
> two threads for this, instead of one doing both, or at least that's how
> FreeBSD does it.
>
>> * does anyone have a plan to update the codebase to recent openzfs?
>
> No, but updating would be great. However, my impression is that the
> zfs world is getting messy and I am not sure that FreeBSD is really
> based on openzfs as clearly as I think you might think. Please just
> take this as a caution that the situation seems more complicated the
> last few years than it was maybe 5y ago.
It would be awful nice if a newer code base could be introduced as I am
sure that there are bugs present in the one NetBSD has and to sync with
some of the newer features. I used to think OpenZFS was the answer but
after reading more on it, I am not so sure any more and it is not clear
to me what a good upstream might be right now.
--
Brad Spencer - brad%anduin.eldar.org@localhost
Home |
Main Index |
Thread Index |
Old Index