First, I'd like to mention I've been in the irc channel (#mac-zfs on freenode) for the past day, and it's sort of fun. Come join us there!
I've been following Don Brady's tweets, which informed me that they're going to be releasing the source for Z-410. No clue yet which pieces and how much of it beyond what's absolutely required, though. I guess we wait and see!
sambo
There's a good interview with Don at <http://arstechnica.com/apple/news/2011/03/how-zfs-is-slowly-making-its-way-to-mac-os-x.ars?utm_source=rss&utm_medium=rss&utm_campaign=rss>
Chris
On Mar 17, 2:47 pm, sammy ominsky <s...@avoidant.org> wrote:I've been following Don Brady's tweets, which informed me that they're going to be releasing the source for Z-410.
Now let's hope when they get their web site together, that they
correct their verbage so that it no longer falsely claims that ZFS on
MacOS is new, novel, or unique!
So the last thing I'd want
to hear from a fellow ZFS on MacOS project, is that our project has
not been freely, generously, and fearlessly advanced for the last few
years by so many valiant volunteers.
It sure looks current!
Dedupe
Compression
Acls
Send/revved
Afp/smb sharing.
-d
-d
-----
Dan Shoop
masterof...@gmail.com
----- Original Message ----
> From: Dan Bethe <smuc...@gmail.com>
> To: zfs-macos <zfs-...@googlegroups.com>
> Sent: Friday, March 25, 2011 11:59:31
> Subject: [zfs-macos] Re: Excellent News from Ten's Complement
>
Ten's Complement is working directly with the Illumos community (the spiritual
successor of OpenSolaris, and core free Solaris code base in the post-Sun era)
to help organize a global ZFS working group. They would help to unify the ZFS
community by providing a reference platform of ZFS, and obviously Ten's
Complement would make MacOS into a top-tier citizen there. They're already
pushing Mac OS-specific patches to them.
Don said that the fact that his company's internal and external behaviors have
appeared to be rather opaque to us, is just an unfortunate coincidence. That's
just been the appearance, not the intention. And he did not see the email CC'd
by Alex. They switched ISPs right at that point in time. His team has had
their heads down, working hard, and taking the fastest, easiest path to
assimilate the latest free ZFS. They regrettably didn't have time for chit-chat
until lately but they do have a public relations person who's developing some
things.
We'll hear more soon, about integrating what he calls "an open source company"
with the community. And people like Alex and Bjoern and Dustin and all the
other contributors and supporters of this community (anyone who's reading this,
especially if you post) have cultured the perfect beta testing base. He lit the
torch and we've carried it.
I think it's easy to say that Ten's Complement has a burgeoning market all sewn
up and ready to be developed. System administrators and prosumers, even just
photographers and other digital artists, are going to be all over ZFS once it's
more fully integrated with the OS and commercially productized and supported. I
would expect the beginnings of a higher level of integration between ZFS and
applications at some point.
It's too early to tell, and I admit that I do like a good hyperbole, but in my
opinion, Mac OS may get its storage strategy back. Call me crazy.
I hope you guys don't mind that I spoke directly for all of us in saying that
Don's our hero right now! I mean, when he decides to.... showwww us the code!
;)
This is the best news I've heard in quite awhile. Thanks for taking
the call for us.
-X
> This is the best news I've heard in quite awhile. Thanks for taking
> the call for us.
Oh my yes. It is my distinct privilege.
I told Don that the MacZFS community is to Ten's Complement like frothing, rabid
geeks camped out in front of the theater, waiting for a release of the next Star
Wars movie.
I hope you don't mind.
Early in my experience, I had asked some people about it, and they said that
MacFUSE was in such disrepair that it was not possible to even start porting ZFS
to it. They say there's only one maintainer, who has been on indefinite hiatus.
I just wanted to ask you all to see if there was a more definitive answer.
Just now, I found this:
http://groups.google.com/group/macfuse/msg/a5e23d8f587993eb?pli=1
And I'm told that this is the most reliable version, where some others are
terrible:
http://www.tuxera.com/mac/macfuse-core-10.5-2.1.9.dmg
So that's all I know!
Well, I am working on connecting MacZFS with MacFUSE. Currently using
MacZFS-74. After porting ztest and in that process libzpool it felt
natural to try and get the zpl part also running in userspace. So far
I'd say it is not particularly difficult, I am just horribly slow in
progress due to very, very limited time to spend on this.
> Just now, I found this:
>
> http://groups.google.com/group/macfuse/msg/a5e23d8f587993eb?pli=1
>
> And I'm told that this is the most reliable version, where some others are
> terrible:
>
> http://www.tuxera.com/mac/macfuse-core-10.5-2.1.9.dmg
>
> So that's all I know!
All the best
Björn
--
| Bjoern Kahl +++ Siegburg +++ Germany |
| "googlelogin@-my-domain-" +++ www.bjoern-kahl.de |
| Languages: German, English, Ancient Latin (a bit :-)) |
> Well, I am working on connecting MacZFS with MacFUSE. Currently using
> MacZFS-74. After porting ztest and in that process libzpool it felt
> natural to try and get the zpl part also running in userspace. So far
> I'd say it is not particularly difficult, I am just horribly slow in
> progress due to very, very limited time to spend on this.
Well that's fascinating. Where is your MacFUSE-ZFS? You're saying that you
ported ztest and libzpool only to MacFUSE-ZFS, not to the kext-based system?
Where can we get it?
If we have the kernel-based MacZFS 74 installed, with a zpool, we don't have
ztest, correct? I haven't seen a ztest binary on MacOS. So there wouldn't be
any point in a non-developer installing your MacFUSE-ZFS just to access ztest on
a production system, because it would only use MacFUSE for all zpool access
rather than the kernel MacZFS, right?
It's the other way around. I ported ztest and libzpool for the
official MacZFS, the kernel-based one mainly maintained by
Alex. I forked his Github repository and made ztest and libzpool
work. Both are in my Github repository. See also Issue #14. Note
that my public Github repository is not up-to-date, the MacZFS-74
version of ztest has not arrived there (had not found the time to
really clean it up). My "port" to MacZFS is not yet in a working
state and therefore not yet released.
Best
Oops, the last line should read:
My "port" of MacZFS to MacFUSE is not yet in a working state and
therefore not yet released.
Björn
Hi there Bjoern. I'm sorry but, just to clarify, it looks like you ported
libzpool and ztest to kernel-based MacZFS 72 which doesn't work on kernel MacZFS
74. Thanks for your efforts.
http://code.google.com/p/maczfs/issues/detail?id=14
But were you saying that you haven't released your MacFUSE-ZFS code at all? I
would imagine that Don might possibly appreciate having what remains of your
MacFUSE-ZFS source code, because it's probably easier to debug a filesystem in
userspace! He did express interest in the topic. Do you think that MacFUSE is
complete enough that it would be feasible to port the latest kernel code to it?
Thanks!
"That was intentional. We took the conservative approach for mounting older
pools (like version 8). There are significant disk changes in the newer pool
format. We simply didn't have the resources to verify the effects of (a)
writing to older pools with the compatibility layer(s) and (b) the effects of
upgrading an older pool. Note that the upgrade process won't upgrade your old
disk format. So you end up with a hybrid pool of sorts. For optimal
performance, it will be better to create a new pool."
-- Don
> The open source strategy of Ten's Complement is still being
> developed. In the future we will announce our plans.
>
> We appreciate the enthusiasm of the MacZFS community, but we're just
> not ready for a wiki since we haven't finished our open source hosting
> strategy.
I'm looking forward to it, even if it is piped through third parties rather than Don posting here (it's not like it's not an open group, after all).
As for needing dedicated infrastructure: if it's going to be open then GitHub is all you need. Most likely, the decision is based oh either making the kext free or paid for, rather than an infra issue. There's also additional GUI tools which might not be opened, which is fair enough.
But right now, it's even less open than it was when it was an official Apple project.
Alex
> As for needing dedicated infrastructure: if it's going to be open then GitHub
>is all you need. Most likely, the decision is based oh either making the kext
>free or paid for, rather than an infra issue. There's also additional GUI tools
>which might not be opened, which is fair enough.
>
Yeah it seems to me that all ya need is a wiki, a web-backed mailing list or
three, a source repository, and a weblog.
> But right now, it's even less open than it was when it was an official Apple
>project.
>
Well. I wonder how long it was going on inside Apple before they got their junk
together enough that the public got anything from it. ^_^
> Well. I wonder how long it was going on inside Apple before they got their
>junk
>
> together enough that the public got anything from it. ^_^
And now this is a far smaller group who is several days behind on responding to
some bug reports. My impression, as they've said, is that they're working as
hard as they can at moving mountains. I have sympathy for em and at this point,
I have a lot of trust (I also trust the law of the CDDL hehe). We also have
people here in this group who are holding on to source code *and* binaries of
interesting stuff for no good reason! ;-)