Important announcement regarding FreeNAS Corral

[Ericloewe]

I wonder if the "jails" problem is really a "warden" problem that will be resolved by FreeNAS 11 moving to iocage.

Iocage is supposed to be much better. I'll reserve judgement until I see it in large-scale action, but it does seem interesting.

Of course, that doesn't solve the "this isn't available for FreeBSD" bit.

 

[indivision]

It is a fundamental function for a storage system. It has no business not being on the GUI.

Which is a great argument for adding it to the GUI. Not canning a massive amount of work that is welcomed by quite a few users.

Most big projects rarely get a real rewrite. The closest was Windows Vista, which significantly overhauled nearly everything, and Microsoft's resources couldn't get it to not be hated.
The normal approach is to slowly review and change components without breaking compatibility, keeping bugs manageable in both scope and number.

FreeNAS isn't anywhere close to operating system big. I've been in software development for 15 years. I can assure you that projects of this scope eventually hit a wall where major portions become dated and need re-writes to make use of new technology.

Even the plans being explained for 9.10 prove this to be true. It includes re-doing the entire UI, changing the underlying OS version, adding middleware, changing the jails system. These amount to a major change also.

It sounds like this is not really about avoiding re-writes. But, about certain developers wanting to work on familiar code. A legitimate issue. But, not the same issue.

It was RC for a day or two. It only got BETA2 when people started complaining.

That is a semantics argument.

I installed and tested a version of what became Corral a long time ago.

It's an example...

...a symptom of a larger problem. If nobody ever hit ESC during development, just what kind of QA happened there? If QA lets the trivial through, what about the less-than-trivial?

That is a subjective, speculative line of thinking to apply to a technical, objective endeavor.

Logically, one oversight does not mean that any other oversights must have occurred or that QA was bad. And you haven't established that it was an oversight rather than a decision. Replacing disks WAS available through the CLI.

Look, there's this fixation on the shiny going on. The shiny is beyond secondary. I rarely interact with the GUI. There's no reason to open the GUI every week, much less every day.

For you maybe. But, that is completely dependent on what the users needs are. I frequently interact with the GUI. And it's not to enjoy how it looks. It's about usability.

I'm pretty sure that there are many FreeNAS users that have non-shiny needs for using the GUI more than once a week.

Most importantly, to everyone saying Corral shouldn't have been canned because code is wasted and such: Have you read the code? The devs have, and they say it's not worth the trouble. That has infinitely more value than the opinion of someone who hasn't even looked at the repository.

This is an open source project. Is it being developed for the community? Or is it being developed for the developers?

Obviously, what Corral delivers is worth the trouble in the eyes of many users. And that should matter. As far as I can tell, there wasn't much of an effort to consider what the community wants from this software before making this change in direction.

You selectively quote and distort my point badly enough that it's hard to believe it's anything other than deliberate, especially when you also quote (and therefore have presumably read) my subsequent posts.

Disk replacement is a critical feature of any redundant storage system--otherwise the redundancy doesn't mean much. For a storage system that's meant to be managed through a web GUI, that functionality must be present in the GUI, and it must be thoroughly tested before release. There's simply no excuse for this feature being absent in a RELEASE version of what's supposed to be a stable storage solution, and that lack, by itself, would (IMO) warrant pulling FN10 from RELEASE status until it's implemented and thoroughly tested--but probably not killing the entire release.

The reason I said that "I guess" killing it was "doing the right thing" was because of the whole history noted throughout this thread. The "don't worry about missing features, it's beta." The "don't worry about this bug, it's beta." The "sure, we'll use an obsolete GUI framework to write our GUI; we'll rewrite it in a different framework in a point release." As I said in the part of my post you didn't quote, it's looking more and more like one bad decision on top of another. I question whether that could be salvaged, but the folks who should be in a much better position to know (i.e., the FreeNAS devs) don't seem to think it can, or at least that it would be worthwhile to do so. And if that's the case, then killing it is the right thing to do.

You claim that I distorted your point. But, then you go on to defend the point as I portrayed it...

I believe that Jordan directly addressed this in a previous post. At some point, software in testing can languish and not progress without the pressure that "RELEASE" labels bring. A "leaving mommy" decision eventually has to be made to cut the apron strings and go release, knowing that there will always be people who take it hard that some certain feature hasn't been added yet. Many people who have submitted feature requests for this project have gone through that ("moved to Nice to Have"...? "Shucks!").

I can agree that disk replacement is important and would ideally be included in the GUI for release. But, I can also understand going release without it and relying on the CLI temporarily.

The bottom line on this point is that claiming this one issue means the whole thing was one bad decision after the other is a hasty generalization.

 

[Dublin]

R.I.P FreeNAS Corral you will missed all i ask with FreeNAS 11 is take your time it took you a year to make 10 i dont care if it takes 2 years all i would say is next time to help trasission over to the next Main update is add new bit slowly even it is for the nightly builds of 9.10.x to help weed out the bugs.

and yes Corral it may not of worked as you wanted it to but this setback will make you stronger as a team.

To all the developers keep up the good work.

 

[indivision]

Because they have a large installed base of users. If they want those users to upgrade, there needs to be a way for them to do so without losing major functionality that iX has long advertised, and they've long been using. "Throw out your jails and replace them from scratch with Docker containers" is the wrong answer. For all the faults of jails (and it may well be that they need to get rid of them in favor of some other technology--the devs seem to think so, anyway), they're out there, in large numbers. And that's because iX gave us the ability, and told us to use it. At one point, it was suggested that a migration path might consist of building a FreeBSD VM, and pulling the jails into that VM--but apparently that idea was dropped in the mad rush to release a half-baked product.

It really is an injustice to describe Corral as a "mad rush to release a half-baked product."

The GUI in Corral is professional level design. In that area alone it is more baked than 9.10 ever was. Not to mention the middleware, docker support, etc. that are all significant achievements that were by-and-large pulled off successfully.

In your comment here, you are basically describing myself a month ago. Switching from jails to docker was an uncomfortable prospect that caused me to put off switching to FreeNAS 10/Corral. And it indeed turned out to be a learning curve. But, it really didn't take that much time. And now that I get how docker works, jails already seem like a distant memory, grouped in with Commando on the Commodore 64.

People are complaining here, very loudly, that Docker support (a feature that's only been released for about a month) is going away with the death of FN10. What of the users of jails, a feature that's been released, promoted, and used for years?

Why do you think people would loudly complain about losing Docker support? Because it is a very good improvement. It greatly expands what you can do with your FreeNAS server (at least without extensive manual CLI work).

The users of jails are limiting themselves by not switching to Docker. Of course, there could be some cases where switching is not practical. And the option always remains to stay on an older build to keep jails.

I'm less concerned with how the GUI looks than I am with whether it works.

Good GUI design is about more than how it looks.

There is a reason why even date-pickers from 2010 do not resemble date-pickers from 2017. There are differences in number of clicks, auto-completing, tabbing, organization, coherency, feedback, etc. that make a big difference beyond just "looking pretty".

 

[Ericloewe]

It includes re-doing the entire UI

So was Corral, for the fourth rewrite or so.

changing the underlying OS version

That's standard BSD procedure.

adding middleware

Much of the work is said to be done. FreeNAS 9.10 can be managed from an API already, at least to a very significant degree.

changing the jails system

No, just the jails manager.

It sounds like this is not really about avoiding re-writes. But, about certain developers wanting to work on familiar code. A legitimate issue. But, not the same issue.

It's about not breaking everything at once. What did work properly in Corral? Samba had some weird issues, replication was only supported to other Corral servers (!), 9pfs needed to be pulled, there was close to zero documentation, the UI was a dead end, important stuff was rewritten and now needs to be independently debugged and fixed (including DHCP)... Just what is going to be salvaged here faster and better than it would be on 9.10?

That is a subjective, speculative line of thinking to apply to a technical, objective endeavor.

Logically, one oversight does not mean that any other oversights must have occurred or that QA was bad. And you haven't established that it was an oversight rather than a decision. Replacing disks WAS available through the CLI.

Ok, so what about 9pfs not actually working reliably? Issues importing encrypted pools? Weird Samba issues? The fact that the so-adored GUI barely works on Chrome and nowhere else and constantly irritates people with expired token messages? No localization support at all? Zero support for screen readers?
This is definitely not good QA.

For you maybe. But, that is completely dependent on what the users needs are. I frequently interact with the GUI. And it's not to enjoy how it looks. It's about usability.

I'm pretty sure that there are many FreeNAS users that have non-shiny needs for using the GUI more than once a week.

Why? The only reason I can think of is "setting up a crapton of FreeNAS servers" and, in that case, muscle memory quickly develops and the UI matters little.

This is an open source project. Is it being developed for the community? Or is it being developed for the developers?

The bottom line is that TrueNAS pays the bills. The community is a nice bonus. That makes it the priority.
That aside, you're not addressing my point. The community, as a large abstract group, is not involved in the dev process and largely has no clue of what is going on in the code. You can't go around saying that it's all roses if you haven't even looked at it, much less tried to make it work and tried to fix it.
If you disagree, go ahead and pick it up. That is the meaning of open-source, not "complain loudly until the developers do exactly what I want even if they feel that it won't work".

I believe that Jordan directly addressed this in a previous post. At some point, software in testing can languish and not progress without the pressure that "RELEASE" labels bring.

Agreed, but that point had clearly not been reached, as is evident from where we are today.

But, I can also understand going release without it and relying on the CLI temporarily.

I cannot. I could manage, but some people would have no clue what to do, especially when faced with a serious lack of documentation. The docs team was explicitly told not to bother until late in the dev process - as preserved in one or two bug reports on the tracker.

Not to mention the middleware, docker support, etc. that are all significant achievements that were by-and-large pulled off successfully.

But were they? None of them seem to work reliably enough for the average user.

Good GUI design is about more than how it looks.

There is a reason why even date-pickers from 2010 do not resemble date-pickers from 2017. There are differences in number of clicks, auto-completing, tabbing, organization, coherency, feedback, etc. that make a big difference beyond just "looking pretty".

You forgot a big one: Accessibility. You're blind? Tough luck, screen readers don't work with the GUI. Oh wait, you need the GUI to be able to setup the CLI for SSH access. Oops, I guess there's no FreeNAS for you. And that's after accessibility was one of the goals for improvement early in the FreeNAS 10 dev process.

 

[userseven]

He does has a point at least: I was always bewildered by the fact that items in left menu and top menu in 9.x did different things despite the fact that they might bear the same caption. Personally what I hate the most about 9.x GUI (functionally) is how it shows messages in an small overlay that fades out in x seconds. Too bad you didn't catch that error in time, now instead of glancing over an immediately accessible event log you have to break out the investigative glasses and find out what (if something) went wrong. Corral got that right, event log should be ALWAYS on screen.

 

[Ericloewe]

He does has a point at least: I was always bewildered by the fact that items in left menu and top menu in 9.x did different things despite the fact that they might bear the same caption.

Yes, that is incredibly silly and I'm glad it's going away. It used to be even worse.

Personally what I hate the most about 9.x GUI (functionally) is how it shows messages in an small overlay that fades out in x seconds.

Yup, pain in the ass.

You just gave me an idea for a feature request, too...

 

[danb35]

It really is an injustice to describe Corral as a "mad rush to release a half-baked product."

No, I don't think it is. I'll admit it's not especially charitable, but I think it's fundamentally accurate. We've seen devs comment on this very thread that there was significant pressure to release, even when they knew of serious problems with the code. There were several examples in the announcement that started this thread of things that just didn't work. The GUI that you're so fond of is fragile af, and that's when using Chrome--with Firefox it's even worse. It's pretty when it works, and it may even make some things easier than the FN9 GUI, but "when it works" is significantly less than 100%. And it was doomed anyway, since iX decided they had to release a GUI written in an obsolete framework, so it would have to be re-re-rewritten in yet another framework after release.

And you haven't established that it was an oversight rather than a decision.

I don't think it does the FN10 devs any favors to assume that the lack of a GUI way to replace a failed disk was a conscious decision, rather than an oversight on their part--either makes them look pretty bad, but I can't decide which is worse. And neither @Ericloewe nor I is arguing that it was necessarily an oversight, and (for my part) I don't think the question is particularly important or interesting.

The users of jails are limiting themselves by not switching to Docker.

Maybe so. So how do we "switch[] to Docker"? As far as I can tell (since iX has given no documentation on the subject), we nuke our jails, find Docker containers for whatever apps we were running in those jails, and install them. Simple enough so far. Then we somehow expose the storage on the FreeNAS box to the Docker container. How do we do that? No clue. No docs. Google it and figure it out. Except that Google probably isn't going to account for the fact that we're running boot2docker in a VM under bhyve on FreeBSD.

You claim that I distorted your point. But, then you go on to defend the point as I portrayed it...

You need to work on your reading comprehension. As I said in what you quoted, but apparently didn't read very closely, disk replacement is (IMO) grounds to pull FN10 from release status until it's implemented and tested. I don't believe there's any excuse for releasing the product without this feature in the GUI. But it isn't, by itself, grounds to kill the entire system--as you say, it can (presumably) be added within the GUI framework. The reason to kill the entire thing is the history of bad decision on top of bad decision, some of which I mentioned in that same block you quoted, and many others of which are discussed elsewhere in this thread, resulting in a codebase which the devs didn't believe could feasibly be salvaged.

Are the devs right in this? I have no idea. They certainly know better than I do, and I expect better than you do too. If you think they're wrong, nothing's stopping you from taking the code and running with it.

 

[raidflex]

For a product to be in such bad shape and then pushed to be released, this shows a seriously lack of senior management in my eyes. All this did was waste the hard work and time of the devs probably just following the direction that they were told to and then feeling pressure from the top to not speak up of all these issues that were present with Corrall.

From the beginning with Docker containers, it should have been made clear that jails would not be able to be imported period. Running a VM (docker) inside another VM on top of an OS only complicates matters and creates unnecessary overhead. It would be one thing if Docker ran nativiley like it does on linux, but this is not the case. Personally I would rather take a little more time to setup a VM for increased reliability and less complexity, then using a container.

Freenas should always be a "NAS" first and foremost. Jails, plugins, Dockers, VMs, etc are great and all but this does not matter if the underlining OS is unstable and the basic functionality of data integrity is compromised. The whole point of using ZFS in the first place was data integrity and the ability to scale. I remember when I switch over to Freenas originally it was such a welcome to move to a platform that did not rely on a hardware raid system and to not have to worry about a failing raid adapter.

 

[JTT0]

Everyone - please cut the team some slack. Yeah, they made a mistake and had some communication issues where they split their teams and did not coordinate with each other about what FreeNAS 10 (Corral) was at release. It happens and is evident that mistakes were made, but the fact that IX came out and owned up to it is phenomenal!

Keep in mind that we have been provided an amazing product FOR FREE by a company that is willing to make sure it is as stable as possible. As I value my data more than anything else on my NAS, I have held off from moving to F10, even though I professionally work as a Software Engineer with Docker containers and love them (I am bringing the functionality to F9.10 via a Jail)! You never upgrade too early, let the kinks get worked out first! With any major rewrite, you can expect a serious number of bugs to show, but in this case it sounds like there were just so many issues under the covers and it was totally unmaintainable (and scary)! As a Software Engineer, I have worked on my share of unstable system cores and I firmly believe in the idea of building major changes on top of a stable core!

IX, please keep up the good work and transparency!

 

[indivision]

So was Corral, for the fourth rewrite or so.

Exactly! So, how is that a knock on Corral in comparison to re-doing the UI again on 9.10?

That's standard BSD procedure.

Exactly! So, switching to 9.10 doesn't prevent the need for big changes, right?

Much of the work is said to be done. FreeNAS 9.10 can be managed from an API already, at least to a very significant degree.

It is said to be done. Is it said to be bug tested? I see no evidence to suggest that bug testing or release will go any smoother for that new work. Why would it?

No, just the jails manager.

What part of the jails system doesn't involve the manager?

It's about not breaking everything at once. What did work properly in Corral? Samba had some weird issues, replication was only supported to other Corral servers (!), 9pfs needed to be pulled, there was close to zero documentation, the UI was a dead end, important stuff was rewritten and now needs to be independently debugged and fixed (including DHCP)... Just what is going to be salvaged here faster and better than it would be on 9.10?

Corral is working great for me. Zero problems with Samba. I upgraded my system to replicate to. Easy. Documentation was being put together at a faster pace than I remember it coming together for any previous build. The UI, even if the framework needed to change, was ahead of 9.10 by nature of it having been designed. It is far easier to apply an existing design with a new framework than it is to re-design AND apply a new framework.

Nothing needs to be "salvaged". Simply prioritize bugs and move forward with them like any other release. Corral is way ahead of 9.10 in terms of architecture, features, design and vision. It does not make sense to step backwards to a system that was already old 2 years ago and kludge features onto it. Maybe that would have been sensible to consider way back when FreeNAS 10 was started. But, now, the heavy lifting is already done in Corral.

I agree with an earlier poster. It sounds like there is more to this story that isn't being put out clearly. Did the developers of Corral get let go (or left on their own)? Is this decision being made primarily because the developers taking over from here know the 9.10 code but do not know the Corral code?

Ok, so what about 9pfs not actually working reliably? Issues importing encrypted pools? Weird Samba issues? The fact that the so-adored GUI barely works on Chrome and nowhere else and constantly irritates people with expired token messages? No localization support at all? Zero support for screen readers?
This is definitely not good QA.

What is 9pfs?

Encrypted pools have always been problematic. In 9.10 also. That's why additional warnings have been given when using them, including in 9.10.

Don't know what samba issues you are seeing. Samba works fine for me.

I use Chrome. Works fine for me on Corral. Seeing A bug report from A person about a browser is anecdotal. You're not seeing the many others who do not have that issue.

Localization is often not included in initial release versions of software. For obvious reasons.

Supporting screen readers (as well as localization) are not QA issues at all. They are decisions.

I find it remarkable that, to some degree, you are representing FreeNAS as a forum mod, while trashing the decisions and effort that FreeNAS has made over the last year+ (as well as the effort that many users pitched in to help troubleshoot, detail feature ideas, etc.). What kind of message are users supposed to take from that?

Why? The only reason I can think of is "setting up a crapton of FreeNAS servers" and, in that case, muscle memory quickly develops and the UI matters little.

Really?

A few reasons:

A) Updating jails/containers.
B) Creating new project spaces/locations/users in a production environment.
C) Giving said users new permissions as needed.
D) Monitoring health of server.
E) Running manual back-ups and other maintenance.
F) Installing new containers/jails. Configuring those. Troubleshooting them as issues come up.

There are many other reasons. It can add up.

Not everyone is just running a fire-and-forget Plex/Downloader.

The bottom line is that TrueNAS pays the bills. The community is a nice bonus. That makes it the priority.
That aside, you're not addressing my point. The community, as a large abstract group, is not involved in the dev process and largely has no clue of what is going on in the code. You can't go around saying that it's all roses if you haven't even looked at it, much less tried to make it work and tried to fix it.
If you disagree, go ahead and pick it up. That is the meaning of open-source, not "complain loudly until the developers do exactly what I want even if they feel that it won't work".

I think that I did address your point as this new comment just brings us back to my same question:

Is the software being developed for users (community or paying) or is it being developed for developers?

I do not need to look at the code in order to have a position on the end product. I am a developer myself. So, if you can find some technical dev explanation for why Corral was bad, I'm pretty sure that I could understand it. To date, nothing like that has been presented. They haven't written that it "won't work". They wrote that it is too much work. Big difference and a legitimate reason for users to be aggravated by after their own investment into Corral (which is work also).

Agreed, but that point had clearly not been reached, as is evident from where we are today.

That is a non-sequitur.

Some people (even those assuming decision-making roles) having the opinion that it wasn't ready is not evidence that it wasn't a good decision.

I cannot. I could manage, but some people would have no clue what to do, especially when faced with a serious lack of documentation. The docs team was explicitly told not to bother until late in the dev process - as preserved in one or two bug reports on the tracker.

So write documentation!

This all demonstrates my point further. The shortest path to the best software is continuing work on Corral. Not reverting back to 9.10.

But were they? None of them seem to work reliably enough for the average user.

They were for me. Obviously. Otherwise, I wouldn't bother writing these posts defending the system.

I do have development experience. But, I'm definitely not above average when it comes to server tech and FreeBSD.

It looks to me like some people didn't want to learn a new system. They want new, cool features, without having to take any time to figure them out. I certainly had to plan time for the transition myself. But, that is just the nature of new software. You can try to offset that by spoon-feeding features (which seems to be the new strategy). But, that doesn't typically result in good software, takes longer to implement new features and in this case wastes a lot of progress that was already made.

You forgot a big one: Accessibility. You're blind? Tough luck, screen readers don't work with the GUI. Oh wait, you need the GUI to be able to setup the CLI for SSH access. Oops, I guess there's no FreeNAS for you. And that's after accessibility was one of the goals for improvement early in the FreeNAS 10 dev process.

Many if not most software does not take accessibility into account in early releases. For obvious reasons.

 

[indivision]

The whole point of using ZFS in the first place was data integrity and the ability to scale. I remember when I switch over to Freenas originally it was such a welcome to move to a platform that did not rely on a hardware raid system and to not have to worry about a failing raid adapter.

I don't see how there was a data integrity difference? You could unplug FreeNAS entirely and start accessing your data with some Linux install instead.

Everyone - please cut the team some slack. Yeah, they made a mistake and had some communication issues where they split their teams and did not coordinate with each other about what FreeNAS 10 (Corral) was at release. It happens and is evident that mistakes were made, but the fact that IX came out and owned up to it is phenomenal!

Keep in mind that we have been provided an amazing product FOR FREE by a company that is willing to make sure it is as stable as possible. As I value my data more than anything else on my NAS, I have held off from moving to F10, even though I professionally work as a Software Engineer with Docker containers and love them (I am bringing the functionality to F9.10 via a Jail)! You never upgrade too early, let the kinks get worked out first! With any major rewrite, you can expect a serious number of bugs to show, but in this case it sounds like there were just so many issues under the covers and it was totally unmaintainable (and scary)! As a Software Engineer, I have worked on my share of unstable system cores and I firmly believe in the idea of building major changes on top of a stable core!

IX, please keep up the good work and transparency!

I appreciate your positivity. But, what team are we supposed to be giving slack to?

What I'm seeing is people throwing others under the bus. I always found Jordan Hubbard's reasoning sound and I see Corral as an achievement. To me, that effort IS what FreeNAS is. What is being described now is a decidedly different effort. I will now have to assess it as such and decide what to do from here. But, as someone who put my own effort into Corral and am very happy with it, it's pretty aggravating to have the rug pulled out from under me.

Regarding never updating too early, that idea renders the "RELEASE" label meaningless. Certainly, practical decisions must be made balancing the likeliness of bugs vs server demands. But, calling software "RELEASE" basically means that the author is saying it's time to download and use it. So, I don't see how anyone can be faulted for not waiting longer.

 

[JTT0]

I appreciate your positivity. But, what team are we supposed to be giving slack to?

What I'm seeing is people throwing others under the bus. I always found Jordan Hubbard's reasoning sound and I see Corral as an achievement. To me, that effort IS what FreeNAS is. What is being described now is a decidedly different effort. I will now have to assess it as such and decide what to do from here. But, as someone who put my own effort into Corral and am very happy with it, it's pretty aggravating to have the rug pulled out from under me.

Regarding never updating too early, that idea renders the "RELEASE" label meaningless. Certainly, practical decisions must be made balancing the likeliness of bugs vs server demands. But, calling software "RELEASE" basically means that the author is saying it's time to download and use it. So, I don't see how anyone can be faulted for not waiting longer.

I am not discounting the effort or reasoning behind decisions made regarding the "release" or the "un-release" of FreeNAS Corral, but simply stating that the team, whoever remains after all the changes at IX, has stated that they have done a deep dive into the code base and determined that the effort to bring F10's major redesign up to required standards is a much longer haul than simply rebasing off of F9.10. This could be due to the remaining team's inexperience with the new systems built or the experimental technologies used for F10, only IX would really know the answer here. In the end, IX made a business decision, hopefully utilizing all of the facts being presented to them and have decided that the huge amount of effort put into F10 was to be picked apart and re-added to F9.10 as time permits.

In any major overhaul of core systems, you will always introduce a variety of critical issues that could potentially impact user data. As I value my data, I was unlikely to move to F10 for some time, even though I really wanted all the new features. As everyone is different and assess the risks involved in their own way, I don't fault anyone for downloading FreeNAS Corral as it became "RELEASE" tagged. In fact, I spent the past month convincing myself not to jump! :-D

Please keep in mind that I totally understand your frustration with the effort put into making F10 the best it could be via bug filing, testing, etc. I commend you for your efforts and time spent. My POV is simply that unless we have delved into the code ourselves (I have not yet), we don't really know what demons lie under that gorgeous UI!

 

[William Grzybowski]

This all demonstrates my point further. The shortest path to the best software is continuing work on Corral. Not reverting back to 9.10.

That is your opinion. I don't agree with but I respect it.
However the decision was been made, the reasons were laid out and thats it.

Can we move on already?

 

[raidflex]

I don't see how there was a data integrity difference? You could unplug FreeNAS entirely and start accessing your data with some Linux install instead.

Issues with drive replacement in the GUI is a data integrity issue as far as I am concerned. This a fundamental part of a NAS, is the ability to easily and reliable replace a faulty drive.

I appreciate your positivity. But, what team are we supposed to be giving slack to?

What I'm seeing is people throwing others under the bus. I always found Jordan Hubbard's reasoning sound and I see Corral as an achievement. To me, that effort IS what FreeNAS is. What is being described now is a decidedly different effort. I will now have to assess it as such and decide what to do from here. But, as someone who put my own effort into Corral and am very happy with it, it's pretty aggravating to have the rug pulled out from under me.

Regarding never updating too early, that idea renders the "RELEASE" label meaningless. Certainly, practical decisions must be made balancing the likeliness of bugs vs server demands. But, calling software "RELEASE" basically means that the author is saying it's time to download and use it. So, I don't see how anyone can be faulted for not waiting longer.

I am not pointing out anyone in particular but in my point of view there would have to be some serious miscommunication issues for a 2-3 year project to just be canned like that. And as I said before I do not think this was the doing of the developers, but a upper level problem. Despite the reasoning behind why Corall was scraped, what is to say this will not happen again. It just leaves a bad taste.

 

[drb]

Can we move on already?

I don't think so. People are going to need to vent. Users of Corral are feeling betrayed. Users of 9.10 are worried about stability and change. Some wondering if iXsystems can recover from this. So let them vent. What harm does that do? Or would you rather they go away?

There is lots at play here within iXsystems and with FreeNAS users. I don't think it is over quite yet. I keep waiting for the other shoe to drop, waiting for an announcement that the former Corral developers have forked Corral and will start work on a community project.

I have no inside knowledge, so this is pure speculation on my part, but the memories of my many years of corporate management politics keeps nagging at me, so I am waiting for something else to happen. I don't think this is over yet.

I wish iXsystems and their former employees all the best in this and I really hope they can all make a soft landing and turn this adversity into a shining star.

 

[danb35]

Many if not most software does not take accessibility into account in early releases.

Perhaps, but "many if not most software" is not written in such a way that accessibility is precluded. The FN10 GUI is.

Localization is often not included in initial release versions of software.

You are again missing the point that the chosen GUI framework precludes localization. These are not simply matters that aren't implemented yet, they are things that can't be implemented within the GUI framework used for FN10.

 

[NeedNAS]

Well I made the jump back to FN9.10. Since I'm still "testing" FN, my box is not live. So albeit the transition back was not seamless, it was also not too difficult either. I set it up to get 9.10 nightly's.
I imported my volumes, got AD up, created a few SMB shares, fixed all the ACLs in the shares and got one VM copied over and running.
But now apparently the FN9.10 nightly's are being being transitioned to FN11 nightly's. My box is in the process of rebooting on the FN11 nightly train now. Fingers crossed.

 

[William Grzybowski]

I don't think so. People are going to need to vent. Users of Corral are feeling betrayed. Users of 9.10 are worried about stability and change. Some wondering if iXsystems can recover from this. So let them vent. What harm does that do? Or would you rather they go away?

There is lots at play here within iXsystems and with FreeNAS users. I don't think it is over quite yet. I keep waiting for the other shoe to drop, waiting for an announcement that the former Corral developers have forked Corral and will start work on a community project.

I have no inside knowledge, so this is pure speculation on my part, but the memories of my many years of corporate management politics keeps nagging at me, so I am waiting for something else to happen. I don't think this is over yet.

I wish iXsystems and their former employees all the best in this and I really hope they can all make a soft landing and turn this adversity into a shining star.

They already have, they still are. One thing is to vent about it, exposing your opinion and wanting answers. Thats is completely understandable and acceptable. Another completely different is trash a company and the developers working on the older product for giving a free product.

A lot of people simply forget iX put a lot of money all these years, and it did not come cheap. That decision did not come lightly, I assure you.

As far as the fork, thats the beauty of open source development. If the Corral code base is so great then I'm sure someone else will keep developing on it.

 

[aim4min]

I, for one, have had no problems whatsoever with Corral. Yes, the migration didn't really work well (I had to start clean) And, I mean, yes, I had to migrate the plex jail to docker. However, I LOVE docker containers. Crashplan never worked right for me before, and the crashplan docker container is amazing. Plus, I was able to quickly set up an OpenVPN container, a UniFi container, and several others. I loved Freenas Corral.

It was a big surprise to find out that is discontinued. This sucks. I don't really want to migrate again. I thought not supporting jails in Corral was surprising, but now, the reverse is true. I really don't want to give up my newfound love for docker.

I liked the new UI too. Had no trouble with it, on chrome, or even on my iphone. (thought i was not super functional - but definitely better than the frames of 9.x)

Anyways, I would rather you spend time fixing the problems, instead of starting over from scratch again. It really wasn't that bad.