This is a really great looking service. I couldn't find a contact form that wasn't for sales inquiries, but I thought you might like to know that your website is extremely buggy on my phone (iPhone 5s, iOS 9.0.2, Mobile Safari).
This is the bug: Scrolling causes the page to reload. Not even joking. Seems to happen on all pages.
Anyway, I took a look on my desktop and it works fine there. I might even sign up a few of my clients. Just not with my phone.
Yes, I did that on my desktop, running Arch Linux.
In retrospect, it probably wasn't a very good idea because my / is on an SSD and ZFS on Linux doesn't support TRIM (I'll probably go with XFS if I ever reinstall), but it's totally doable.
Sure you can, Proxmox has distributed ZFS with it's custom Debian-based OS for over a year now.
ZFS comes pre-installed and the installer for Proxmox even gives you the options to install the OS directly onto a bootable zpool that is created automatically by the installer itself.
The thing I miss most in ZFS (aside from a native in-kernel linux port, as I've found ZoL to be kinda buggy) is to be able to grow zdevs. I can add a new zdev, I can remove them, but I can't add one disk to my 7-disk RAIDZ2. MD RAID has no problem with this.
Also, if you add zdevs later, you will never get full performance again as the striping will be off (putting more writes on the zdevs with more free space), and there's no 'rebalance' kind of thing (I've seen some scripts that just move stuff around, but I've never seen any evidence that they work, and I really doubt that they do based on my understanding of zdev allocation).
I'm assuming you mean vdevs — you can resize zvols (virtual block devices) just fine.
(and you can even grow disk-type vdevs, though they'll be somewhat unbalanced)
Me too. I've added so many drives in onesies and twosies to so many MD RAID5/6 arrays over the years that I simply can't confidently deploy ZFS for that case.
Array reshapes are rare, and it'll only use idle IO bandwidth anyway. Redundancy is guaranteed during the whole process. So I don't see this as a problem.
> BTRFS supports adding a new drive, reshape (balance) is optional
Thanks for pointing this out. I didn't know btrfs supported array reshapes. I'm waiting for it to mature (RAID5/6 was only added in 3.19, for example) but I expect I'll switch over at some point, probably at the point it's considered "boring" (exactly the experience I want from my filesystems).
Both are to blame. Neither are to blame. Much of the success of Linux has been attributed to the GPL. It certainly had a role to play in opening up WRT firmware.
The truth is that the GPL is a pretty restrictive license. I don't think it's fair to blame anyone for the fact that it's not compatible with much. That's its defining feature and the main reason people choose it; so if it blocks ZoL, it's working as intended.
Which I consider a great loss. ZFS is fantastic.
I use the BSD 2-clause license on my own stuff, but I can think of situations where I'd rather use GPL.
There seems to be a ton of commercial and proprietary software and services on Linux with the GPL. Too many people just don't understand the GPL.
If you make changes you must share the changes does not equal Plague Like.
GPL is the most used and misunderstood license. It is more complicated now than I wish it was but it has given birth to so much. I prefer it to BSD or MIT due to the share a like nature of GPL.
> If you make changes you must share the changes does not equal Plague Like.
You don't even have to share the changes with the unwashed masses, or even the original author, just the people you sell/give the resulting program/service too. Of course they are free to distribute the changes more freely, the license explicitly states it is a breach for you to try block that, but you don't have to.
The GPL does not require you to release your modified version, or any part of it. You are free to make modifications and use them privately, without ever releasing them. This applies to organizations (including companies), too; an organization can make a modified version and use it internally without ever releasing it outside the organization.
But if you release the modified version to the public in some way, the GPL requires you to make the modified source code available to the program's users, under the GPL.
Thus, the GPL gives permission to release the modified program in certain ways, and not in other ways; but the decision of whether to release it is up to you.
Though in the "a company or other organisation" situation IIRC your internal users count in the same way so far as the GPL is concerned so you should make the code available to them on request if you don't already and they are free to redistribute under the terms of the GPL. Another important caveat is the "upon request" part: you don't have to release the source to anyone unless explicitly requested so you don't have to release it automatically upon releasing a binary.
Of course if you just use the results for your own needs/entertainment there is no requirement to release anything at all.
I'm no stranger to the GPL. I had a project which spent 9 months arguing with legal in a large institution to support 6 months work involving an independent open source contractor to work on a GPL'd thing.
But it is a restrictive license. Its "plague"-ness comes from its incompatibility with other licenses. Or at least, in the case of ZoL/CDDL and Linux/GPL, the crippling practical realities of the resulting kernel module binaries that are a derived work trying to impose restrictions on each other. That's why you can find precompiled modules on the ZoL website, but never in a distro.
That's a legitimate criticism, but as you say, also its defining feature.
> Much of the success of Linux has been attributed to the GPL.
That I know, much of the success of Linux has been attributed to GNU, and mostly in the sense that it made Linux's legal background very clear while BSD's libc was in legal limbo (Linux used glibc which had been funded by the FSF for the express purpose of having a clearly free software libc)
Couldn't they have dual licensed it under both the CDDL and GPL? CDDL for when they distribute it with the closed drivers, then additionally distribute the source under the GPL allowing it to be used in Linux.
Listen to it and the whole point he is making is we had to make way for proprietary drivers and not to block it being compatible with GPL.
Seem very much a grammar Judo to say we didn't go out of the way to be anti-compatible with GPL but we invented a copy left license so we could never be compatible with GPL.
Well how much work can you expect a for-profit-company to do in order to release their software as open source? They choose a licence that worked for them, then and there because of reasons. This included the work on ZFS.
There are operating systems that have licences that permit the ZFS code in the kernel. What reasons did SUN have to specifically block the Linux kernel but not BSD / Darwin ? The choice of license has everything to do whit what license SUN was comfortable using and nothing to do with keeping GNU/Linux out of the party.
If the answer is that a lot of their competitors at the time were invested in GNU/Linux and thus could not use their code, that is a very short sighted move.
[1] http://www.rsync.net/products/zfsintro.html