wishing Hammer was ported to OpenBSD

[Oko]

Originally Posted by matiasbsd

No virtualization for OpenBSD :-(

OpenBSD is network appliance OS. There are many much more important things than stupid Virtualization which are missing on OpenBSD. You want to do some permanent good for our community. Why don't you port a serious file system like Hammer to OpenBSD

[ibara]

I don't really like this approach at all.

Originally Posted by Oko

OpenBSD is network appliance OS.

No... OpenBSD is a general purpose operating system. Sure, there are some things that OpenBSD excels at and other things that it can improve upon, but let's try to avoid shoehorning. You will miss out on the amazing things OpenBSD can be if you let your imagination sing.

I use OpenBSD everyday for everything - my router, my fileserver, my desktop, my game center, my schoolwork center, my multimedia experience, my social media hub, my jukebox, everything. In fact, the thing I probably use it least of all as is network appliance.

Originally Posted by Oko

There are many much more important things than stupid Virtualization which are missing on OpenBSD. You want to do some permanent good for our community. Why don't you port a serious file system like Hammer to OpenBSD

Virtualization is a fact of life in 2014. I for one would love to see someone interested in getting good virtualization host support on OpenBSD. This is something OpenBSD should obviously excel at: a secure, sane virtualization host. But it's not my itch.

And yes, HAMMER2 would be nice as well.

[Oko]

Originally Posted by ibara

I use OpenBSD everyday for my schoolwork.

I am sure you do

Originally Posted by ibara

And yes, HAMMER2 would be nice as well.

When you get in my age you will learn not to work with imaginary software or file systems for that mater. Folks at the moment of writing this post HAMMER2 (with exception of tiny pieces) doesn't exist in reality except in the head of Matt Dillon. Consequently it can't be ported to OpenBSD. Porting Hammer to OpenBSD which actually exists would be really nice.

[ocicat]

Specific musings of porting the Hammer filesystem has been separated from the parent thread:

http://daemonforums.org/showthread.php?t=6960

...as the original discussion centered on the lack of virtualization solutions available today.

We ask members to respect the direction taken by the initial poster creating each thread. Side discussions warrant their own threads as this simplifies the arc of discussion, & it makes archive searches simpler.

[ibara]

Originally Posted by Oko

I am sure you do

I am decades older than you think I am; spare me your sarcasm.

Originally Posted by Oko

When you get in my age you will learn not to work with imaginary software or file systems for that mater. Folks at the moment of writing this post HAMMER2 (with exception of tiny pieces) doesn't exist in reality except in the head of Matt Dillon. Consequently it can't be ported to OpenBSD. Porting Hammer to OpenBSD which actually exists would be really nice.

The opening statement contains what is called an ad hominem attack (succinctly: you accuse me of being a child, therefore I do not know how things work), and it is incredibly stupid on your part considering you do not know who I am. The irony of course is that which you accuse me of is an exact match to your behavior. To save yourself further embarrassment, please refrain from such attacks in the future.

Now, onto adult conversation:
HAMMER1 is very tightly integrated into DragonFly, so much so that untangling HAMMER1 from DFly is probably too much of a task. HAMMER2 can still be brought in piece by piece: a much less herculean task.

Since you're the one who brought it up and since you clearly did not understand my reply, I'll make it more direct: why don't you start porting HAMMER1 to OpenBSD? That's even more "permanent good" than telling someone else to. I'll even personally review your code. Please reread this post. It applies to things in base too, not just ports.

[Oko]

Originally Posted by ibara

Since you're the one who brought it up and since you clearly did not understand my reply, I'll make it more direct: why don't you start porting HAMMER1 to OpenBSD? That's even more "permanent good" than telling someone else to. I'll even personally review your code.

I am not asking anybody to do anything

[J65nko]

I cannot find anything to substantiate it, but IIRC the Bitrig developers are/were considering Hammer.

[Oko]

Originally Posted by J65nko

I cannot find anything to substantiate it, but IIRC the Bitrig developers are/were considering Hammer.

They haven't made any progress just like they have not made any progress with FUSE. However in the mean time OpenBSD got FUSE as we know and it was promptly ported back to Bitrig. Fuse is the first step in porting HAMMER to OpenBSD. Now waiting for HAMMER2 is a valid strategy as I see that many DragonFly users consider HAMMER only a legacy file system. Whether that is justified or not we will see. Matt is definitely one of the most capable world programmers as we know. Surprisingly enough many serious Linux users are existed about HAMMER which they see as a counter balance to ZFS. I am guessing many of those users are no more delusional and realize that btrfs
doesn't exist in reality and might never exist as so many things started but never finished in GNU camp.

[backrow]

Originally Posted by Oko

They haven't made any progress just like they have not made any progress with FUSE. However in the mean time OpenBSD got FUSE as we know and it was promptly ported back to Bitrig. Fuse is the first step in porting HAMMER to OpenBSD.

Is it? How useful would it be in userspace? (it’s a serious question—I don’t know how HAMMER works.)

FWIW, there is a Google Summer of Code proposal for a student to do get started with porting HAMMER to OpenBSD (or rather a viability study). If you know students capable of that or anything else on the list, think about convincing them to apply.

[PrinceCruise]

Originally Posted by Oko

Surprisingly enough many serious Linux users are existed about HAMMER which they see as a counter balance to ZFS. I am guessing many of those users are no more delusional and realize that btrfs
doesn't exist in reality and might never exist as so many things started but never finished in GNU camp.

I don't want to speak on their behalf, but Linux users are neither excited about HAMMER nor had any issues with ZFS to see it as an imbalance. ZFSOnLinux port wasn't created for no reason.
And are you saying that btrfs doesn't really hold any value for Linux users even after 5 years of continuous and persistent development? I agree it's still not considered "production-ready" because its in continuous development but it is marked stable. SUSE guys ain't mad, RHEL guys ain't mad and the testers/end-users who spend their time improving btrfs by using it and reporting any issues ain't mad either.

Regards.

[thirdm]

Originally Posted by PrinceCruise

I don't want to speak on their behalf, but Linux users are neither excited about HAMMER nor had any issues with ZFS to see it as an imbalance. ZFSOnLinux port wasn't created for no reason.
Regards.

This is closer to my experience with Linux users. I was the only Linux user (used slackware for a while last year) I ever met who wanted hammer. I don't think many Linux users know much about it other than as something like btrfs or ZFS that runs on something called DragonFlyBSD, and you're lucky if they've even heard of DragonFly. Justin Sherrill's done a great job with his news summaries in publicizing DragonFly by being compelling even to non-DragonFly users, but my feeling is that he hasn't yet broken out beyond BSD users. (Not to say that's even his goal, but it could happen eventually I'd guess. There just aren't that many really good tech. news aggregators.)

What I've also come across is lack of interest in any FS in this category among home users and hobbiests, the feeling being that these features are only relevant to servers. I've also seen less interest in ZFS than in btrfs and simultaneous slagging of btrfs.

I second the sentiment that it would be really cool to see hammer (I or II) ported to other BSDs, particularly OpenBSD.

[PrinceCruise]

Originally Posted by thirdm

This is closer to my experience with Linux users. I was the only Linux user (used slackware for a while last year) I ever met who wanted hammer. I don't think many Linux users know much about it other than as something like btrfs or ZFS that runs on something called DragonFlyBSD, and you're lucky if they've even heard of DragonFly. Justin Sherrill's done a great job with his news summaries in publicizing DragonFly by being compelling even to non-DragonFly users, but my feeling is that he hasn't yet broken out beyond BSD users. (Not to say that's even his goal, but it could happen eventually I'd guess. There just aren't that many really good tech. news aggregators.)

Slackware user here too but I had my eyes on DragonflyBSD/HAMMER since the day I knew about it.

What I've also come across is lack of interest in any FS in this category among home users and hobbiests, the feeling being that these features are only relevant to servers. I've also seen less interest in ZFS than in btrfs and simultaneous slagging of btrfs.

This.
While being a part of the discussions in various Linux/tech forums, I realized that a lot of Linux users who are aware of the HAMMER file system are not really excited about it mostly because for a home user it doesn't matter much to use so much advanced features of hammer, ZFS or btrfs, if only they do is watch movies and browse online.
And in server space, most of them and even I'd be cautious to come out of my comfort zone of ext* or xfs filesystems.

Regards.

[Jtf]

Originally Posted by Oko

Why don't you port a serious file system like Hammer to OpenBSD

Hi Oko, I happen to have found this GitHub repository which is related to what you were saying.

Unfortunately, there doesn't seem on be much activity on there. The last commit was 4 months ago.