Which is more important to you, unRAID 5.0 features or 3TB drive support?

[neilt0]

I see a big factor in this issue being the "company" behind the product.

 

AFAICT, it's not a company, it's one person. One person who developed the software to cater for his personal needs, then decided to sell it.

 

Once you start selling something, you need to cater for the needs of the purchasers, assuming they are reasonable and common amongst many users.

But if you are a one-man band, it's often the case that you ignore feature requests if you don't agree with them or don't see them as a priority. Even if the majority of users are requesting them.

 

I've come across this with the new version of Newsbin, which I've been using since about 1999. V6.0 has removed two key features that I and others rely on and the developer is refusing to enable them, despite it being very simple, because "we're doing it wrong" by not using the new automated features he's provided.

He actually said "I don't use that feature, so I'm not going to put it back in".

 

That's fine, as long as you don't expect people to buy the software. His response was "If you don't like the new version, use the old one", which is unrealistic and childish.

 

This means I no longer use the program I purchased. I have had to switch to other (ironically) free apps.

 

There are a few features in unRAID that have been requested which don't seem to be a priority. I really don't understand not having the option of a Recycle Bin or similar. An accidental press of the Delete button can wipe out 40TB of data. I know I can run a telnet command to (maybe) unerase the data, but not having a Recycle Bin is a real pain. It means I have to unnecessarily set most of my shares to Read Only to avoid those accidents. On more than one occasion, I've accidentally deleted a folder with over 40GB of data.

 

Not having someone on Limetech here at least once a day to answer the unanswered questions is unacceptable for a product that costs up to $149 (and which is what I paid).

[chuck23322]

I did just buy unRAID the other day.  However, if I had known beforehand it didn't support 3TB drives I certainly would NOT have purchased.  Just my 2c.  I am still happy to support the project but 3TB support will soon become an issue with prospective buyers if it isn't added.  TBH I never would've guessed that an OS released in 2011 wouldn't support >2TB.  In the computer software world when you face such fundamental limitations and you still haven't overcome said limitations when the hardware is out in the wild, you've dropped the ball.  I know that sounds harsh but it's the truth.

 

You mean, the free version of unRaid was available to you to play and test the limitations of it?   And you didn't catch the limitations then?  And the forum was online when you were doing the testing, and you didn't notice the limitations there?   And the roadmap for 5.0 was online, and you didn't catch the limitations there?

 

Not sure why you're complaining, but that's OK.    I'm a noob here with unRaid,  and couldn't be more happier with it.  Even with the current 2tb limit.  Heck, most OS's and MoBos can't deal with 3TB anyway.   And that's just bleeding edge.   Most people would take a rock solid unRaid server than an extra terabyte per drive.  And it really isn't a huge hassle for most of us to decide to buy a 3tb drive now, knowing it's in the roadmap to use that extra terabyte later.  {After lots of testing that your MoBo, controller, etc are up to the job}

[SSD]

Heck, most OS's and MoBos can't deal with 3TB anyway.   ...

 

This just isn't true.  Please do your homework before posting such statements as fact.  Linux does support the 3T drives (as does Windows 7).  Virtually every motherboard out there will support 3T drives, as will most add-on controllers.  These drives have been on the market for 6 months or more, and are well tested and supported.

 

My opinions aside on what should be the unRAID priorirty, let's not confuse fact with opinion.

[lionelhutz]

This needs to go to beta and then everyone will know what support issues there are,

Haha, and then you get ALL the support emails and requests with customers complaining about why it does not work.  I am not saying you are wrong, just stating a fact. 

 

An unRAID fairy isn't going to magically appear before Limetech one day and announce "3T drives will now work perfectly so go ahead with support". Limetech needs to create 3T support that works in house and then release that as a beta so others can test it too. Otherwise, the support will never be developed and 3T drives will never be supported.

 

The arguement that it's a feature that might not work so it shouldn't be supported yet is complete crap. You can use that same arguement against every possible new feature and never go anywhere with development.

 

Peter

 

[bubbaQ]

The arguement that it's a feature that might not work so it shouldn't be supported yet is complete crap.

 

Not at all.  Unless *you* are the person dealing with the support issues, it is definitely not "crap."  And it is not a "might" but a "will".  There are controllers that do NOT support the 3TB drives, and the compatibility of many others is unknown.  AFAIK, for maximum 3TB reliability, you need to be using AHCI, and how many people either 1) don't have a mobo that supports AHCI, or 2) don't have it enabled in the BIOS?

 

unRAID will support 3TB drives at some point, just as it will support other hardware advances eventually.  All this bitching and moaning is about timing, not implementation, and it is human nature that people want their particular wheel greased first.  I'm not bothered by 6 months more before 3TB support is shipping, and if I was putting on a project manager's hat, I'd feel the same way.  If I was a user maxed out on 2TB drives, I'd be unhappy not with unRAID, but with the fact that I am in that tiny minority of the market that I need something that is not yet available.

 

People forget that unRAID is a storage appliance, and not some marginally useful pet application that will not cause great havoc if it pukes.  I have some experience with code for an implantable medical device, and the degree of code review and testing the FDA requires is Herculean---but that is expected since it deals with human life (note that the MS license agreement at one time expressly declined use in any system on which human life depended, for liability reasons).

 

unRAID isn't a medical device, but when you are dealing with years worth of someone's data, it is pretty close second in terms of importance to a lot of people.  It is a heck of a lot higher on the "risk avoidance" totem pole than most any other applications.  All development decisions that can affect reliability of data should be made with the utmost of caution.

 

The drive manufactures have taken the same approach.  They do not expose 4K sectors to the OS, because some applications might choke, and the drive manufacturers will get blamed.  They are shipping their own controllers, because so many controllers currently in use are not safe to use with 3TB drives.  I don't see anyone harpooning the drive makers for these decisions, despite the fact they reduce performance and limit the consumer's use of the products.

[neilt0]

I have some experience with code for an implantable medical device.

 

[boof]

unRAID will support 3TB drives at some point, just as it will support other hardware advances eventually.  All this bitching and moaning is about timing, not implementation, and it is human nature that people want their particular wheel greased first.  I'm not bothered by 6 months more before 3TB support is shipping, and if I was putting on a project manager's hat, I'd feel the same way.  If I was a user maxed out on 2TB drives, I'd be unhappy not with unRAID, but with the fact that I am in that tiny minority of the market that I need something that is not yet available.

 

Where is this 6 month figure coming from? Would you bothered by 7 months more? What do you know we don't?

 

And timing is everything. There is no point having a storage appliance that doesn't support current tech. Or do Drobo etc who are commercially providing storage appliances with 3TB disk support have all this wrong as well?

 

3TB drives are available on the market. Other storage appliances (both hardware and software based) do already support them. So no, I as a user would be unhappy that unraid is now lagging behind.

 

unRAID isn't a medical device, but when you are dealing with years worth of someone's data, it is pretty close second in terms of importance to a lot of people.  It is a heck of a lot higher on the "risk avoidance" totem pole than most any other applications.  All development decisions that can affect reliability of data should be made with the utmost of caution.

 

No it's not a medical device so I have no idea why you needed to drop that in. If we go by the above then you would never include any new features. And this is, as has been pointed out, what the beta release cycle is for.

 

And none of this changes the fact we have no firm roadmap or date for 3TB support. As far we should be concerned at the moment it is never. Again you may know something we don't.

 

The drive manufactures have taken the same approach.  They do not expose 4K sectors to the OS, because some applications might choke, and the drive manufacturers will get blamed.  They are shipping their own controllers, because so many controllers currently in use are not safe to use with 3TB drives.  I don't see anyone harpooning the drive makers for these decisions, despite the fact they reduce performance and limit the consumer's use of the products.

 

This would be a better argument than 'it's hard so unraid should be cautious in implementing'. Are you saying that current controllers out there that are recognising and putting data onto 3TB drives right now are playing fast and loose and aren't safe to use?

 

If that's the case and you have data to support then it would indeed put a whole new spin on things.

 

As a random aside (given we're dropping random things into posts) if you, as a project manager on a commercial project chose to de-prioritise development of a feature that your direct competitors had already brought to market which gave them a large selling point over your own product - you'd be very very different from any project manager I've ever worked with.

[lionelhutz]

The arguement that it's a feature that might not work so it shouldn't be supported yet is complete crap.

 

Not at all.  Unless *you* are the person dealing with the support issues, it is definitely not "crap."  And it is not a "might" but a "will".  There are controllers that do NOT support the 3TB drives, and the compatibility of many others is unknown.  AFAIK, for maximum 3TB reliability, you need to be using AHCI, and how many people either 1) don't have a mobo that supports AHCI, or 2) don't have it enabled in the BIOS?

 

By using your arguement - unRAID should never have been developed at all, since there are SATA controllers and motherboards which do not support it. Sorry, but bitching and moaning about possible support issues is complete crap. No-one knows how good or bad it will be until it's implimented.

 

The best way is with a beta release to test the support and find issues. If it really tanks then 3T support can always be removed for the final release.

 

Peter

 

[kizer]

Just keep in mind everybody this poll was setup just as a feeler to see what everybody thinks. It wasn't created by limetech. It was created by a forum member as a moderator just to keep discussion going. Its results does not currently drive development one way or another. Limetech could either gauge what members have voted as a decision or completely ignore because it doesn't make sense.

 

Nobody knows the direction or when 3TB will be implemented. Some might argue thats exactly the problem and others like my self will simply roll with it until it is.

[aiden]

Part of the problem here is this topic has touched a nerve with a large number of us who are simply tired of being in the dark and tired of sporadic communication from the developer.  Out of that frustration we attack each other because that's what humans do when they are upset and irritated.  The only real solution is for Lime-tech to take a long hard look at its current operations model, and make some drastic changes.  I was excited at the beginning of the year with the new forum realignment, the Roadmap, and the new mod positions, and it had all the trappings of such a drastic change.  However, it didn't last very long and we are all back to playing the "someday" game, with some power users arguing about why 3TB support is unimportant or irrelevant to their needs and that we all just need to sit back and wait.  The user community for unRAID, in my opinion, is expected to have the patience of Job. 

[bubbaQ]

Are you saying that current controllers out there that are recognizing and putting data onto 3TB drives right now are playing fast and loose and aren't safe to use?

 

Yup.  One of the manufactures of 3TB drives listed dozens of controllers in some documentation linked in a previous thread, identifying good/bad ones.

 

Why do you think WD is bundling a controller with the 3TB drives?

[boof]

Yup.  One of the manufactures of 3TB drives listed dozens of controllers in some documentation linked in a previous thread, identifying good/bad ones.

 

Have you a link to the thread? And this also means there *are* ok controllers...

 

Why do you think WD is bundling a controller with the 3TB drives?

 

I had presumed it was in case you had a controller that would not pick up the drive correctly full stop. Rather than a controller that would find the drive with no problems and write data to it with problems (?).

 

Hitachi don't bundle a controller with theirs... (edit : I take that back, they do if you don't buy OEM)

 

Any comment on the 6 month time span you linked to 3TB support? As a global moderator of this forum you're much closer to limetech than any of us - so when you throw a number against something like that we have to assume you have better information than us and are speaking with more authority....

[prostuff1]

Any comment on the 6 month time span you linked to 3TB support? As a global moderator of this forum you're much closer to limetech than any of us - so when you throw a number against something like that we have to assume you have better information than us and are speaking with more authority....

 

That was when we would like to see it added by, as far as having more information than you... that is not entirely true and in this instance we have no more information than you do.

[bubbaQ]

if you, as a project manager on a commercial project chose to de-prioritise development of a feature that your direct competitors had already brought to market which gave them a large selling point over your own product - you'd be very very different from any project manager I've ever worked with.

 

I never said de-prioritize.  I said I would evaluate the positives and negatives regarding the timing of the release, the negatives including the risk to users' data from hardware that I can not control.  The negatives also including the support headaches and blame received for something that was not my software's fault.  I would certainly prioritize the development --- but the release would be weighed against the negatives.  Release today might get a thumbs down, while release in 6 months might get a thumbs up.  It is a balancing act.  I don't see other storage appliances rushing out to copy unRAID's features, so perhaps their managers are incompetent too?

 

We can debate whether 3TB support is a "large selling point" today or not, but no one would convince the other.  There are not enough resources (time, money, etc.) to satisfy everyone so the developer has to decide what is appropriate, and that is guaranteed to mean different priorities from those of at least some of his customers.  I don't think any of us is in a position to tell Tom what to do with his product.  After all, he is in a much superior position to evaluate his market than any of us are.

 

Don't get me wrong --- I would love to see 3TB come out tomorrow if Tom is up to it.  But how many people have full 2TB arrays?  Not too many.  I'll wager that the vast majority of inquiries about 3TB support are from people enamored of the latest-great-new-shiny-thing syndrome.  I had a engineering firm client I kept on Novel 3.11/3.12 for many years, and several times they a manager always wanting to upgrade to version XXXX.  I counseled against it, and prevailed for several years.  Then she did it anyway, and it broke their production job management system.  They had no need to upgrade, but the latest-great-new-shiny-thing syndrome was just too strong to resist.

 

Managing a data storage appliance project has different priorities than managing development of something like a browser or a word processor.  Different projects have different issues.  Each project's risk versus reward balance is different.

 

IMHO, a 20:1 data:parity ratio (40TB to 2TB of parity) is way too risky.  From an data integrity perspective, if I was managing a storage appliance project, I would direct the vast majority of resources to diagonal parity.  I wouldn't release 3TB support until I had DP working.

[boof]

Don't get me wrong --- I would love to see 3TB come out tomorrow if Tom is up to it.  But how many people have full 2TB arrays?  Not too many. 

 

You're making assumptions.

 

There is also the question of the TB per port cost when taking into account the infrastructure needed in the server. For me 3TB is sweeter now than 2TB. Which makes adding any disks that aren't 3TB 'more expensive' then it needs to be.

 

As you say, not much point in continuing to argue but I am now interested in what happens in 6 months if there's no further movement on 3TB support...!

[aiden]

IMHO, a 20:1 data:parity ratio (40TB to 2TB of parity) is way too risky.  From an data integrity perspective, if I was managing a storage appliance project, I would direct the vast majority of resources to diagonal parity.  I wouldn't release 3TB support until I had DP working.

 

And IMHO, a 33% reduction in disk quantity mitigates to a large extent the risk associated with a 20:1 array by increasing the size of each disk by 50%.  It's two sides of the same coin, but I can't imagine >2.1TB support being more complex than diagonal parity.

 

As you said, everyone has their opinion, with valid arguments, and valid concerns.  Does anyone else feel like this thread is chasing its tail?

[kizer]

Funny side note.

 

2Months ago a friend bought a bright and shiny new HP from Best Buy. Good Choice or bad choice I don't know since I would of built my own, but anyways.........

 

Dropped in a new 3TB from NewEgg and it did come with a controller. Bios didn't like the 3TB so it downgraded the drive to something smaller. I then tried the controller card and it saw the 3TB, but windows did not. I did some research via my phone since she didn't have internet and from what I gathered she needed to have Windows7 64bit for it to work. Sure enough she did, but no matter what I did it woudn't see over 2.2TB

 

Maybe I'm a moron, maybe I should of spent some more time on it, but in the end right out of the box the 3TB from WD did not work. Does that mean my current unRAID build will/wont work? Beats me, but honestly I'd hate for people to really get all worked up about wanting 3TB support if/when they find out what they are running might not even support it anyways. I don't know of a qualifying list anywhere that says what will and what will not.

[BRiT]

It's far more entertaining to watch this heated discussion from the sidelines than trying to educate some folks who are making obviously wrong assumptions based on lack of knowledge. 

[Josh]

I'm fully populated in my 4224 and am a month away from needing to build a new machine.  Im trying to hold out until 3tb support. I'd like to be done with 2tb drives because in the long run they will have to be replaced with something bigger and I'd like to avoid that.  Start with 3's and then move into 4's is my goal, doesn't look like that will happen though

 

Please tell me you've posted in pimp my rig thread.  We always need more impressive shots.  Any yours sounds like it's got bragging rights!

 

http://lime-technology.com/forum/index.php?topic=10085.0

[VampyreGTX]

Wow, unbelievable on how heated this has become! I view it as such, I have a Norco 4220, and could run up to 40TB of data in the current iteration, with 3tb, I could expand that to 60TB of data. BUT, would I want that with a possibly unstable 5.0? Let's get 5.0 up, let's ensure it's stable through a first, full build, and then start looking at 3TB drives. That is a LOT of data that can be lost. I don't feel like re-ripping 100+ BR's cause I have a failure and lose an entire drive or two of data. But, that's just me.

[SSD]

Wow, unbelievable on how heated this has become! I view it as such, I have a Norco 4220, and could run up to 40TB of data in the current iteration, with 3tb, I could expand that to 60TB of data. BUT, would I want that with a possibly unstable 5.0? Let's get 5.0 up, let's ensure it's stable through a first, full build, and then start looking at 3TB drives. That is a LOT of data that can be lost. I don't feel like re-ripping 100+ BR's cause I have a failure and lose an entire drive or two of data. But, that's just me.

 

Are you saying that if Tom implements 3T drive support you are going to lose your data?

 

ANY new feature can have bugs - even the simplest ones.  This enhancement would not be a big one.  It is the OS, after all, that writes data to your data disks.  unRAID just has to maintain parity, and once the GPT paritions are in place, unRAID will do just that.  And it is very easy to test.  Just run parity checks and disk rebuilds. 

 

Scaring yourself or others into believing that 3T support means data loss is pretty silly.

[Auggie]

AFP support is important to Mac users

AD support is important to Windows users

Greater than 2.1 TB support is going to be imporant to everyone, especially to new users and the competition.

 

if this is an overly simplified summary of what to include in 5.0 release, then I would say that AFP support is not critical since SMB and NFS are both functional for Mac users (I being one of them).  They are not the perfect solutions for us Mac users, but they function and for the moment, it's good enough so long as AFP support will eventually be included in a future release (5.1?).

 

As for AD support, my opinion is that this is mostly a business/organization application (the exceptional "power user" that implements this in their home network would comprise a small niche), which should be a very important group of customers to cater to.

 

Greater than 2.1 TB support should DEFINITELY be on the 5.0 release list.  With WD 3TB's now going for $150 on Amazon (and these still include the HBA adapter; imagine what the price would be without out) the ROI on 3TB drives is much better than with 2TB drives, extremely important for customers beginning brand new builds, and customers that have or will soon reach maximum expansion of physical drives.  Those customers in the latter group (which I am) are now facing a real financial and physical burden due to the lack of 3TB support.  The only current option would be to build a new, separate server but that means purchasing the lesser ROI 2TB drives, allocating physical space for the new machine as well as higher maintenance and power costs associated with it.  For enterprise users, this may not be an issue, but for smaller businesses or home users, this option may not be feasible.

[ragnar]

given i can get 4TB for $160 Vs 3TB for $199 i'm not to worried about 3tb drive support yet.

 

it will probally take at least another year before 3TB drives reach anything like 2tb drive prices so i think all the other benifits are worth it first.

[Wimpie]

given i can get 4TB for $160 Vs 3TB for $199 i'm not to worried about 3tb drive support yet.

 

it will probally take at least another year before 3TB drives reach anything like 2tb drive prices so i think all the other benifits are worth it first.

 

And where do you plug those 2Tb drives in? In 0$ SATA ports?

 

You have to include these costs also!

 

If your case is full, will you buy a complete new server for 0$?

 

 

For people who have an (almost) full server, 3Tb support is extremely important!

 

Even for new users, who have limited driveslots (eg max 6 drives), it is extremely important that they can now use a 3 Tb drive. Otherwise they have to buy a 2Tb drive that they later have to replace with a 3Tb one.

 

6 months ago, 3Tb support was maybe not so important, but 6 months is an very long time in the IT world. A lot has changed since, and I think it is now very important to have 3Tb support as soon as possible, before AD, AFP and  other not so essential things.

 

Just my 2 cents...

 

 

 

 

[PeterB]

Otherwise they have to buy a 2Tb drive that they later have to replace with a 3Tb one.

 

No, they can buy a 3TB drive now and get an immediate 10% increase in capacity over a 2TB drive.  Then, when unRAID support does arrive, they get a further 36% increase for no additional cost.

 

Of course, you should all count yourselves lucky.  I consider myself fortunate when I can find a 1TB drive on the shelf, a 2TB drive is a special order, with a 2 week lead time, and I don't think that any shops here even know that 3TB drives exist.