It feels like a new beta is imminent, thoughts on what it might contain?

[JustinChase]

So, the announcements board has been changed to not allow us to post, so I guess this is where the discussion will happen from now on.

 

jonp said last week (Wednesday?) that he expected to have the new beta out by the end of the week (a week ago), but we know that's usually a bit optimistic, and historically has taken a week or 2 beyond his guesstimate.

 

A lot of people complain about it, but it doesn't bother me.  I'm really happy there is so much more and better communication than there was for the first couple years I was coming here.

 

jonp said in another recent post that a new release was coming soon, and they just made the new posts for the IRC channel, and closed the announcements board, which I assume are pre-cursors to the new release.

 

So, what do we think will be included?

 

Personally, I think we'll see email notifications, smart reports, and more VM tweaks.

 

I hope we'll see soon.

[WeeboTech]

Knowing Jon, we'll surely see some kool   VM/Docker functionality.

 

I can tell you guys that I've been lobbying hard for visible smart data from the webGui.

This is and syslog data is crucial to the first level of support when hard drive or potential corruption issues show up.

 

With rsyslog, we can have some form of automated message filters and events, however, given the schedule, I'm not sure about automated alerts yet.

 

I don't have any more information then you think, other then I've been pushing hard to get smart data visible for the common folk and for us to support them.

 

First level is visibility. I have not pushed hard on attribute monitoring... YET! LOL.  We may be surprised!

[garycase]

I have no idea what is actually going to make it into the next Beta, but based on the emphasis on virtualization support and the amount of dialogue in that direction I'd expect to see significant improvements in that area (although the Docker emphasis may detract a bit from that for now).

 

I'd hope to also see notifications and UPS support -- I suspect we'll see that by v6 final, but not sure about the next Beta.

 

Not sure there's much point in speculation, however, as the release is surely CLOSE ... although the fact it wasn't out last week likely means there were some last minute "hiccups" -- and there's no telling how long those might take to resolve.

 

 

[BRiT]

I'll be a bit more pragmatic.

 

I doubt we will see a release in October or even the first weekend of November. I doubt the next beta will contain anything that is "quick and easy" to implement that should be fundamental core NAS functionality such as alerts, email alerts, drive health, smart reports, ups support, or schedule health and system checks.

 

Instead it will contain noncore NAS functionality like VMs. And it will contain some annoying bugs.

[WeeboTech]

To be fair, October is over.

 

In addition nothing is quick and easy. For a single admin on a couple servers, perhaps.

When you are dealing with an OS that goes out to many people of various hardware and custom configurations, core NAS changes have to be really solid.

 

Should these fundamental core NAS features be implemented. Heck yeah.

As a paying customer, reach out and let your voice be heard.

 

Respectable communication in Roadmap forum is Key.

http://lime-technology.com/forum/index.php?board=63.0

 

If it's something that is really important to you, I might suggest reaching out directly to limetech. I have a number of times.

[BRiT]

In addition nothing is quick and easy.

 

I'm not the one who used those terms; those were LimeTech's words when they explained why they were going after the virtualization functions first and foremost.

[reggierat]

Anyone have a log of the chat earlier, had to head off and do some shopping

[kaiguy]

So I read through the IRC chat log.  Not so sure that's the best way to disseminate information, but I guess it's better than nothing.

 

My takeaway is this, and sorry if it's overly simplified: Dynamix is now officially part of unRAID, and there's going to be more GPU passthrough stuff... So, officially add-in existing capabilities, and a way to use a huge multi-drive server as a living room client.  Am I really that out of touch that I don't see how this will be a popular use case?

 

I've asked this before, and I'll keep on asking: what is the longer-term vision of unRAID? Because I honestly don't understand why a lot of these development decisions are being made/prioritized.  It feels like an identity crisis.

[WeeboTech]

and a way to use a huge multi-drive server as a living room client.

 

I don't see a 'huge' multidrive server in the living room as a common usage.

However I have seen other products that are co-located in a closet in a Home theater setting whereby the gui is propagated to the home theater and around the home.

 

The merging of file storage and graphical front end is not so uncommon.  I've done it with a smaller XBMC arrangement with 4 built in drives.

 

 

The other product I've seen, and repaired, has been retired and the company dismantled.

At that time it ran 3ware in raid 0 with mythtv as the front end on Fedora core 2.

You could rip dvds to storage and then view them with myth TV.

The guy loved it so much I couldn't talk him out of going with individual machines.

All he cared about was this consolidated machine and he just wanted to throw money at fixing it.

 

I think unRAID's potential to fill this niche market is good. Maybe not for a 24 drive server but for an average home user with small array needs it's attractive. Those who require larger arrays may not need the marriage of reliable storage and gui front end in 1.

 

[kaiguy]

I think unRAID's potential to fill this niche market is good. Maybe not for a 24 drive server but for an average home user with small array needs it's attractive. Those who require larger arrays may not need the marriage of reliable storage and gui front end in 1.

Fair enough.  I guess I'm somewhat unRAID old school.  I want to keep the server/client relationship separate.  Perhaps thats the Plex user in me, where I like to have thin clients displaying my media, and not have my server try to be a jack-of-all-trades.

 

I just wish unRAID core functionality didn't keep getting a backseat to these new use cases.

[WeeboTech]

I just wish unRAID core functionality didn't keep getting a backseat to these new use cases.

 

 

I think dynamix adds some of the missing core features we need.

[garycase]

It sounds like an interesting evolution ... but I'll withhold judgment until I see it.  I also tend to be a bit old-school UnRAID ... I want a rock-solid storage appliance that includes the core NAS functionality that's been missing for so many years.    If it happens to support more, that's okay, but NOT if there is ANY impact on reliability.

 

[jonp]

I've asked this before, and I'll keep on asking: what is the longer-term vision of unRAID?

 

Big Picture

What previously was an OS to give you complete control over your storage is now an OS to give you complete control over your hardware in general (CPU, disk, network, RAM, and other devices) as well as your software (applications and virtual machines).

 

A new solution to an old problem...

The primary reason NAS solutions have remained as headless appliances has been due to software limitations.  Simply put, application developers tend not to write their software with other applications in mind.  They write the software with the OS in mind, but the presence of other applications in the same OS can, at times, cause conflicts that result in system instability.  Sometimes this is due to resource contention in hardware (CPU, RAM, Disk, Network) and sometimes this is due to software dependency conflicts (aka dependency hell).  In fact, for a long time, NAS solutions didn't even let you install ANY software on them.  They would simply provide a few protocols and basic management interface.  If you wanted apps, you were told to buy a general purpose server and put them there.

 

Traditional NAS solutions have tried to solve this problem and provide a home for apps by creating unique ways to isolate applications on the same OS (think Jails).  But even these approaches come with limits.  The center of the problem stems from the belief that applications have to "live" within the NAS OS itself.  With Docker and VMs, we can give each application a home to live in both independent from each other as well as from our NAS OS.  This allows safe and secure ways for users to install not only traditional server applications but full blown desktop applications as well.  And not just for our OS, but any OS you want.  Here's an example:

 

 

This is my setup at the lab.  I change the inputs on the displays for various configurations, but here's what I have running in this picture:

 

On the left, I have my macbook air and a secondary display connected over VGA -> Thunderbolt above it.  In the center, I have Windows 8.1 running as a virtual machine, but I am passing through a GeForce GTX 780.  On the right of that, I have an Ubuntu 14.04.1 VM running with pass through to a much older AMD Radeon card (HD4350).  I have one mouse and keyboard to control all 3 of these simultaneously.  I am running Synergy (a software client) that allows me to set up a massive multi monitor display across various devices with various operating systems.  Now do I think this is an atypical user setup?  No, not at all, but it shows the flexibility of the technology.  While I work, I have Time Machine backups occurring from my macbook to the unRAID server.  I also have Docker containers running for Plex and several other apps.  I can leverage AFP, SMB, NFS, and FTP all at the same time.

 

Also, I have been capturing lots of video footage lately in preparation for getting more guides out there to folks in a more efficient manner.  Know how I've been capturing footage?  I have a capture card in that unRAID server passed through to the Windows VM.  I then map a network drive from Windows to unRAID and have all the video footage I capture write straight to the array.  I have literally gone HDMI out of my AMD card to the HDMI in on my capture card.

 

So let's go over a few potential use cases for all this stuff:

 

1)  If you have a PC today that has capable hardware, you could turn that PC in to a NAS and not give up the "PC" part.  This way, when you want to upgrade your hardware, you don't have to rebuild your desktop as it's all the same virtual hardware to the OS.

 

2)  If your a gamer, but you hate all the noise generated by the loud fans in your rig, you could put your beefy card in the NAS, put it out of sight, then use Steam in-home streaming to get the same quality graphics to any other system in your house (even laptops) without the noise.

 

3)  If you have a gaming PC and a NAS, you could leave your gaming PC as-is, but still add a low-end GPU to your NAS, attach USB devices for control and again use in-home streaming to stream your games to your NAS with the same quality graphics as your PC.

 

4)  If you're a videographer or photographer, you could eliminate the network layer when you want to copy data to your NAS by using a VM with pass through.  Attach your equipment directly to the unRAID server and perform a local copy.

 

5)  If you're an IT pro or developer that wants to perform really fast dev/test work or lab work, you could build a large # of VMs in a cache pool on SSDs for quick build-up/tear-down testing.

 

6)  If you want "true" desktop virtualization, you can have one PC for multiple individual users each with their own monitor, mouse, and keyboard, and eliminate the use of remote graphics over a network connection that leads to such a poor user experience.

 

And this barely scratches the surface.  You'll see far more potential use cases from us as we have the opportunity to prove them out and showcase them on our forums and blog.

 

What if I have a low-end system?

An item on our roadmap that we have yet to implement is an additional boot mode (we call it "unRAID Lite").  This will drop all the packages from the build that are related to VMs so if you have no interest in Xen or KVM and just want a basic NAS and maybe a few Docker containers, you can do that.  This build will probably be about half the size in memory of the current builds as well, allowing you to run unRAID on even less memory than you can today.  I have transcoded videos with Plex using a Celeron processor to give you an idea of the efficiency you can achieve.

 

What about core NAS features?

The merger of Dynamix to the mainline build is the first logical step towards integrating the core NAS features everyone's been asking for into unRAID.  Notifications, SMART data reporting, and more on their way along with plenty of other features.  Dynamix was just the first step too.  We plan on incorporating other community works where relevant and necessary.  The reality is that the core features needed to be added, but they weren't going to be our points of differentiation from competition that already has those features.  Dynamix helped us play some major catch-up and added some other nice advantages (real-time reporting, dashboard interface, etc.), but virtualization and Docker pushes us way ahead of the rest of the solutions out there in terms of providing a solid solution for application support.

 

Wrapping it up

Not everyone is going to use all the capabilities we bake into unRAID.  The reality is that we want to build an operating system that can scale with the user's hardware and better take advantage of all it's capabilities.  Many folks may start off a build that is still a low-end headless NAS, but as time goes on, folks upgrade their hardware, and unRAID supports that just fine.  But we also think a lot of folks have built some pretty beefy systems to support their NAS, and we want to show them ways to fully leverage all of those capabilities.

 

[WeeboTech]

Great work Jon.  BTW, I think you should scale down that image before posting it (or repost a smaller version). Most people dont have 3264x2448 screens.

 

Re: What if I have a low-end system?

An item on our roadmap that we have yet to implement is an additional boot mode (we call it "unRAID Lite").

 

This is why I brought up the ramfs/rootfs vs tmpfs conversation in the IRC.  with tmpfs and a swap file it can help with low memory systems.

ramfs cannot be swapped, tmpfs can. While memory is affordable today. swapping out files or pages that are hardly used allows the rest of ram to be used effectively.

[jonp]

Great work Jon.  BTW, I think you should scale down that image before posting it (or repost a smaller version). Most people dont have 3264x2448 screens.

 

Done!  Thanks for the suggestion.  I posted from phone so it scale automatically in tapatalk.  Didn't realize that didn't occur through SMF.

[jumperalex]

also: your ... you're ... please oh god please 

[JaY_III]

Jonp can you please post that in announcements. If that is your vision for unRAID, you don't want it to get buried in this thread

 

[JustinChase]

Not as imminent as I thought/hoped it seems.

 

I have a theory on this.

 

I suspect that LimeTech feels like they are ready for a release, then discover some bug report on the forums, then find a 'quick fix' and then run the fixed program thru all their rigorous testing, which delays the release.  I suspect this happens 'regularly', which is why they don't put out releases as often as they like or commit to.

 

Now, I don't know if that's at all accurate, but assuming it is; I don't think this is the 'right' way to go about things.  If they are ready to release, and something comes up, it's almost certainly something that has already existed in the program previously, so it's not going to 'hurt' anything to leave the 'bug' in the program a bit longer.  And by leaving it, they can release the updates more regularly, and let us try/test the updates that are ready, and they can fix the bugs, and let the release with the fixes come out later, in a new/different release. 

 

This would

let us 'play' with new things more frequently

let us test the new features on a wider mix of hardware

keep some folks from 'complaining' that we don't see updates often enough (or that promises are not kept)

enhance the feel of a more regularly updated/enhanced product.

 

Just my $.02

[WeeboTech]

Not as imminent as I thought/hoped it seems.

 

I have a theory on this.

 

I suspect that LimeTech feels like they are ready for a release, then discover some bug report on the forums, then find a 'quick fix' and then run the fixed program thru all their rigorous testing, which delays the release.  I suspect this happens 'regularly', which is why they don't put out releases as often as they like or commit to.

 

Now, I don't know if that's at all accurate, but assuming it is; I don't think this is the 'right' way to go about things.  If they are ready to release, and something comes up, it's almost certainly something that has already existed in the program previously, so it's not going to 'hurt' anything to leave the 'bug' in the program a bit longer.  And by leaving it, they can release the updates more regularly, and let us try/test the updates that are ready, and they can fix the bugs, and let the release with the fixes come out later, in a new/different release. 

 

This would

let us 'play' with new things more frequently

let us test the new features on a wider mix of hardware

keep some folks from 'complaining' that we don't see updates often enough (or that promises are not kept)

enhance the feel of a more regularly updated/enhanced product.

 

Just my $.02

 

 

I don't think that's a good approach, Every issue will have a trail/thread of communication and that takes time.

Squash what you can, document what is outstanding, but put out what you believe is a reliable release.

 

 

None of this half-fast stuff!!!

 

 

[CUclimber]

Jonp can you please post that in announcements. If that is your vision for unRAID, you don't want it to get buried in this thread

This bears repeating.  I mean, they have a blog on the main LimeTech site but the announcement for the next version came in the form of an IRC transcript .txt file (seriously?!) and now we get this great post buried in a thread.  Use the blog!  That's what they're for!

 

That said: this is all good news.  Despite its current drawbacks I'm still an UnRaid fan and I'm glad that we will be getting more functionality in the upcoming release.  As soon as I'm comfortable with a backup system for it (my laptop needs rock-solid Time Machine capability to the server, and I need easy folder-level offsite backup to Crashplan/Backblaze/etc.) I'll start playing with VMs and whatnot, which will be great to have in one centralized place on my home network.

[dalben]

The one constant of unRAID is the absolute inability to hit a delivery date.  In fact, it's so much a defining attribute that if they were to ever deliver on a date they forecast more than a month out, I would wait a lot longer to install it than I usually would for fear that their usual due diligence was compromised.

[archedraft]

Lol

[Thornwood]

If dynamix is added I am so excited to have this on v6 again I can hardly wait.

[jonp]

If dynamix is added I am so excited to have this on v6 again I can hardly wait.

It is ;-)

[JaY_III]

Lime technology and valve must be using the same clock. https://developer.valvesoftware.com/wiki/Valve_Time

Now that I know this, the roadmap and updates of soon make more sense