For fun: what stops unRAID being a major contender


NAS

Recommended Posts

Excellent!

 

Does the API come in 4.5, or do we wait until 5.x?

 

 

It will be 5.X, but I will try to get a 5.0-beta1 out soon and in parallel with 4.5-beta series.

Way cool...  Thanks Tom.

 

With an defined API the unRAID product will really be able to mature.  (I expect a lot of fun is in store for some of us)

 

Joe L.

Link to comment
  • Replies 97
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

In the interface changes I have been mulling, where it insulates the user from having to do anything "Linux", I've tried things that things that have eye-candy appeal as well as functionality.

 

Once of these is some system monitoring, which includes CPU temps, and even fan speeds.

 

Would you think about enabling Hardware Monitoring in the kernel, with the modules for AMD and Intel temp sensors, and the it87?  AMD and Intel will make CPU temp available for essentially everyone, and the it87 is very common. 

Link to comment

While a lot of people want a prettier GUI (me included to an extent) I think the main point we are getting at is that we want an open API that we can hook into.  That way the talented developers around here can go wild and design there own interfaces.

 

Sprucing up the default GUI is not a bad thing to do but i would much rather have stuff done under the hood.  Let things like unRAID-Web have the hooks it needs and we can take the GUI from there and make it what we would like.

Link to comment

Would you think about enabling Hardware Monitoring in the kernel, with the modules for AMD and Intel temp sensors, and the it87?  AMD and Intel will make CPU temp available for essentially everyone, and the it87 is very common. 

 

Sure this can be added.  Here is the list of h/w monitoring chips supported by the kernel, which of these do you suggest?

 

--- Hardware Monitoring support
< >   Abit uGuru (rev 1 & 2) (NEW)
< >   Abit uGuru (rev 3) (NEW)
< >   AMD Athlon64/FX or Opteron temperature sensor (NEW)
< >   FB-DIMM AMB temperature sensor on Intel 5000 series chipsets (NEW)
< >   Fintek F71805F/FG, F71806F/FG and F71872F/FG (NEW)
< >   Fintek F71882FG and F71883FG (NEW)
< >   Intel Core (2) Duo/Solo temperature sensor (NEW)
< >   ITE IT87xx and compatibles (NEW)
< >   National Semiconductor PC87360 family (NEW)
< >   National Semiconductor PC87427 (NEW)
< >   Silicon Integrated Systems Corp. SiS5595
< >   SMSC LPC47M10x and compatibles (NEW)
< >   SMSC LPC47B397-NC (NEW)
< >   VIA686A (NEW)
< >   VIA VT1211 (NEW)
< >   VIA VT8231 (NEW)
< >   Winbond W83627HF, W83627THF, W83637HF, W83687THF, W83697HF (NEW)
< >   Winbond W83627EHF/DHG (NEW)
< >   IBM Hard Drive Active Protection System (hdaps) (NEW)
< >   STMicroeletronics LIS3LV02Dx three-axis digital accelerometer (NEW)
< >   Apple SMC (Motion sensor, light sensor, keyboard backlight) (NEW)

Link to comment

These are the must-haves:

<>   AMD Athlon64/FX or Opteron temperature sensor (NEW)

<>   Intel Core (2) Duo/Solo temperature sensor (NEW)

<>   ITE IT87xx and compatibles (NEW)

 

Winbond are also pretty common... :

<>   Winbond W83627HF, W83627THF, W83637HF, W83687THF, W83697HF (NEW)

<>   Winbond W83627EHF/DHG (NEW)

 

I have those enabled on all my kernel builds.  

 

It looks like the .28 kernel combined several individual modules together (i.e. there used to be separate k7 and k8 AMD modules).  That makes things a lot cleaner.

 

Link to comment

First, I completely appreciate all the comments, observations, and suggestions, both positive and negative.  I never thought unRAID would generate such a following, even if it's "I love to hate it"  :)

 

I think this thread boils down to two main gripes:

1. The GUI is ugly.

2. We want more functionality.

 

So here's the plan.  First, I will be releasing the 4.5 beta series & 4.5 will be the last release with the GUI in it's current closed form.  Next, I want to start on version 5.0.  This version will include lighttpd and php, with the current GUI pages converted to php.  More importantly I want to "open" the API so that anyone can add additional functionality.  It is abundantly clear that there are some very talented people in the unRAID community & I think everyone could benefit greatly by "turning them loose".

 

I'm late to the game in this thread, but this is a huge jump for unRAID. Extremely wise move.

FWIW, I don't think the gui is ugly.. But I sure would like to add some of my own functionality easier.

Thanks for listening. (reading?)   ;D

Link to comment

These are the must-haves:

<>   AMD Athlon64/FX or Opteron temperature sensor (NEW)

<>   Intel Core (2) Duo/Solo temperature sensor (NEW)

<>   ITE IT87xx and compatibles (NEW)

 

Winbond are also pretty common... :

<>   Winbond W83627HF, W83627THF, W83637HF, W83687THF, W83697HF (NEW)

<>   Winbond W83627EHF/DHG (NEW)

 

 

OK, added these 5 to the kernel build so you'll see it in 4.5-beta3.

Link to comment

Since you have the hood up, would you consider adding lsof to the initramfs?  Now that we have a more dynamic interface, it is really nice to be able to give the user a lists of open files, particularly when shutting down, which along with fuser, can make shutdowns "cleaner."

Link to comment

Thanks for the reply, we all know you're busy.

 

This feature has been available almost since day 1 in the GUI.

 

Being told that, I'd guess now that it must the "Disk shares (SMB)" setting... which wasn't clear to me by looking at it, or even by reading the unRAID manual, partly because because it doesn't explicitly explain why the SMB is a drop down and the NFS is a text box.  Having two different controls for what is ostensibly the same purpose is confusing.

 

How many media server environments use LDAP??  However, we are releasing Active Directory support.

 

Well, I guess there's at least one here, though I guess this is probably about as common as the need for, say, Subversion and an internal PKI infrastructure, which I also run.  I guess I'll hold off on any request for support for client+server authentication via certificates (like what VMWare 2 and Terminal Server do) for now.  ;)

 

As a corollary, I find it hard to explain in detail to my g/f why our condo needs 8 machines (and she doesn't even see all the VMs).  ;)

 

The 'security model' is not designed to let anyone other than a single admin (called root) "log in".  User level security is facilitated by the Admin specifying the users permitted or not permitted to access shares.  The fact you have to create the users (and passwords) locally is a necessary artifact of Windows networking.  Aforementioned Active Directory support will alleviate this for many environments.

 

Sounds good... that'll probably fix what I'm looking for.  I believe it's fundamental networking security for users to be able to set their own passwords for a system they are accessing, even if admins control what they can access.  Whether users are permitted to access something is authorization, whereas login itself is authentication; these are different.

 

Not true, User Shares are available in all versions, always have been.

 

I don't know if I was drinking too much eggnog, but I seem to remember something involving user shares not working when I tried the trial (user level security?).  I was only on the trial for 2 days anyway, since, like I said, I knew I wanted the product after evaluating others (the process for which I stated here).

 

One thing I've learned working at commercial software companies is that a lot of time customers just don't want the product first for no particular reason (maybe they just don't want to spend the money, or they don't see it as a necessity), and then they'll tell you what it's missing as the reason they don't buy it.  "Well, if it just did X, I'd buy it" they say.  Then you put X in and come back to them, and they say "well, if it just did Y, I'd buy it".  Reality: they just don't want it.

True indeed.  Also they say, "add feature X and you will sell many more of them"... right.

 

Well, if it picked the winning lottery numbers reliably, it's probably sell a lot, and people wouldn't even complain about the serial being tied to the USB key.  ;)

 

Anyway, I love unRAID, but that's because I knew what to expect going in and ultimately I wanted the product.  I basically agree with bjp999's last post... I see unRAID as a niche product for those who need to economically store a lot of data with some level of protection, no more.

This indeed was the original product spec.

 

I think unRAID has more than succeeded at this.

 

If I needed automated backups or plug and play install, obviously I'd have to look elsewhere,

Automated backup yes, Plug and Play - what do you mean by that?  Pre-defined shares?

 

At this point, I have no idea what I was thinking.  :-[  I may have just meant a GUI style install with auto detection of hardware, with no need to run funny commands to prep a USB key (as simple as that is for most semi-technical users).  Obviously, Lime sells UnRAID servers just as Dell sells WHS machines, so maybe this wasn't valid.

 

but until it makes economic sense for Lime Technology to hire dozens of engineers to boost the product, they should not get into the habit of catering to the "well, if it just had DoD level encryption, I'd buy it" crowd.

Hard learned lesson.

 

In that case, I think you're smarter than about 98% of software development managers I've met and/or had the displeasure of working for.  :D

 

IMO, I think the best way to find where demand will lie is to just look at where users are going and technology is shifting.  Trends I'd look at would be social networking/Web 2.0 (probably not applicable to unRAID but maybe someone can think of something), ultra-mobile computing (ditto... netbooks, iPhone, etc), cloud computing (distributed encoding anyone?), the advent of SSDs, rapidly falling HDD prices (perhaps being influenced by SSDs), and continued acceptance of broadband and media content-over-IP.  Set the roadmap to that and stay ahead of the users, rather than chasing features that reflect trailing technology.

 

So aside from UI improvements, the features/functionality I'd be most looking for myself would be:

1. Expansion of the number of drives and second parity, based on the principle of readily available inexpensive HDDs (I know this was slated)

2. Optimization of I/O for SSDs... i.e. make sure SSDs have proper partition alignment (if necessary- maybe nothing is needed here)

 

Love the open API concept.  IMO, after converting the current GUI for 5.x, just let the community design GUIs and don't even bother with .css in the default unRAID GUI.  I was considering writing a Windows Desktop gadget myself (don't like keeping a browser window open all the time), and the first step would have been to have written an abstraction layer that ran in its own process that I'd communicate with.  Heck, I still might just because I'm so old school I don't like the idea of client polling to get server status.

 

Plus I think there's potentially a lot more to GUI that just server statuses.  As a geek I like to see my syslog and check on my free space and all, but if I really want to make my life better it'd be nice to have some kind of media browsing tool that runs on my HTPC that is somehow "unRAID smart".  So I could have an unRAID on the backend serving up media and a friendly front end that is more than a just a media server browser like most apps.  In other words, as an admin, I might simply place content on the unRAID server; front end users who switch to the HTPC channel on my HDTV would suddenly see this as "newly published content" and use the remote, air mouse or whatever Wii controller like thingy to watch it.  To be truly rich, it might require a facilitator of some kind on the unRAID side.  Maybe even a "push" feature that silently caches files at the HTPC itself in case the network bandwidth isn't there for streaming playback.  I know there are things like Popcorn Hour but nothing is quite at the level of "what can my retired tech-phobic mom use?"

 

And I don't know how controversial this would be, but as a registered user of unRAID, I don't mind paying for a upgrade for major revisions.  I know that's not as good as $0, but at some level I think those of us who want to make requests should put their money where their mouth is as well.  :D

 

For now, though I have a job and my lunch break is up. Maybe if I end up unemployed, I'll implement some of this stuff, because in the end it's 10000x easier to me and others talk about features and GUIs than it is to actually do work, which is why I definitely respect Lime & those in the community who contribute.  Thanks to everyone for the discussion... but especially to those who put all the work that goes into unRAID and its addons.

 

Link to comment

Since you have the hood up, would you consider adding lsof to the initramfs?  Now that we have a more dynamic interface, it is really nice to be able to give the user a lists of open files, particularly when shutting down, which along with fuser, can make shutdowns "cleaner."

 

Yes, now added to 4.5-beta3.

Link to comment

I'll make another point.

 

Currently at my house I have Asterisk running my phones. My phone bill is $10/month and the system rivals most Office systems.

I have LCD's placed in several rooms streaming music, videos, emails, recipes in the kitchen etc etc.

I also have home automation controllers sequencing blinds, doors, lights, plugs, motors etc.

I have Unraid handling all of my backups and media.

 

Now. You're all probably guessing the first thing I do is lead dinner party guests into the server storage room and show off Unraid's GUI. Well, surprise! WRONG!

 

If someone asks about the storage server, I MIGHT explain to them how well Unraid works at protecting my collection of movies and how easy it is to increase the storage pool. But that's as far as it usually goes until their eyes start glazing over. I end with, "Its the most cost-effective/price performance data server I've found on the market today."

 

If there's any iPhone Drones at the party I'll show them my movie collection with Poster/Fan Art in XBMC. That usually keeps their attention for more than 5 minutes.

 

So Tom, please for the sake of the product. Keep performance,reliability, compatibility, and sensible features at the forefront of your development.

 

Leave the fancy GUI work for any left over time you have. GUI Geeks [pronounced Goo-e geeks] can stay busy with their iPhones.

 

 

Link to comment

keep in mind how much dev time were talking about as well.

 

People are saying "not no fancy GUI if it gets in the way off.... X".

 

But 10 minutes a week of dev time to date would have been more time than BubbaQ has had to do the entire lighty project.

 

Its not as simple as saying doing X takes away from Y.

 

Anyway this is all old hat now. With V5 all things change.

Link to comment

keep in mind how much dev time were talking about as well.

 

People are saying "not no fancy GUI if it gets in the way off.... X".

 

But 10 minutes a week of dev time to date would have been more time than BubbaQ has had to do the entire lighty project.

 

Its not as simple as saying doing X takes away from Y.

 

Anyway this is all old hat now. With V5 all things change.

 

No offense, but this is pretty funny.  :D :D :D

 

There is no way to work on something like this for 10 minutes a week and make any progress at all.

 

And even if you added all of those 10 minutes a week together and spent three years of time all at once, you're only talking about 26 hours.  That ain't nothin'!  :o

 

Developing a modern, aesthetically pleasing GUI for unRAID using graphics and modern Web design, would consume Tom for quite a while. and absolutely take time away from other initiatives and enhancements.

Link to comment

[

No offense, but this is pretty funny.  :D :D :D

 

There is no way to work on something like this for 10 minutes a week and make any progress at all.

 

developing is not magic... you spend time you make progress. no idea how things work in your company but I can happily do 10 mins on something and make progress. nit picking on math is just being silly.

 

anyway your flogging a dead horse the thread achieved its purpose. since its conception we have a new lighty community GUI and confirmation of a new version coming from the developers.

 

Im glad i started it and put up with all the "i know better than you talk :)"

 

I think it best I close this thread now cause i can see no good coming of nit picking like that.

 

Well done all. Especially well done Tom and Kudos to BubbaQ for striding out and making stuff happen.

 

Locked

 

 

 

 

 

Link to comment
Guest
This topic is now closed to further replies.