I've been using unRAID for a few months now, experimenting in different setups, using different plugins and different virtualization setups (standalone, virtualbox plugin, under ESXi, under Xenserver...). I've also been comparing unRAID to other home/soho NAS solutions, to find the best fit for my needs.

I love unRAID, and will probably stick with it. Its unique features make a lot of sense and its stable, solid architecture is a plus as well.

 

With all that, there are some features that I miss. I thought I'd list them here. They have (obviously) been mentioned in the past in these forums. Perhaps other people would add their "top two" into the thread, in hope to provide some useful feedback to Limetech.

 

1. Integrated encryption support

 

Filesystem- or device-based encryption layer, that will live somewhere below the unRAID share features layer. It could conceivably be under the unRAID md, or, better off IMHO, on top of md but under the disk and user shares. The most natural candidate for this is dm-crypt and LUKS. Ideally, I'll be able to configure it from the GUI, and it will take a passkey, a keyfile or both to open the container. The container could be a file (in which case the share on top of its mount point could be a user share) or a whole disk. Due to the nature of unRAID it probably can't be the whole array (which, if it were possible, would have been ideal).

 

Yes, I can use truecrypt today (see my entry here). But with that, I miss all the nice unRAID features at the share level etc.

 

An interesting question would be how to deal with the cache.

 

2. Dual parity (aka P+Q, aka RAID6-style, aka...)

 

As mentioned in quite a few other posts here, double drive failure is not uncommon and not as unlikely, statistically, as would initially seem. One failure plus subsequent rebuild can drive a second failure before rebuild completes.

Murphy works 24x7.

Calculating a second parity would minimize the damage from such failure. To me, this seems quite important.

 

I said top two, so the above are mine 

 

Way further down on my list there's another feature that would be nice to have, which I mention here only since I don't remember seeing it elsewhere on the forums:

Better clean/unclean indication on the array

Currently, it seems that every time unRAID shuts down uncleanly (e.g. power fail), on the next boot it will assume the array is dirty and parity check will be required. It appears as if even when the array is idle at the time of the fail, and definitely not writing anything, the assumption will hold.

It would have been nice if the array were to contain a dirty bit ("write in progress") which will be reset on p-write completion. Then, when the array is brought online, this bit can be examined to determine cleanliness, and rebuild would not be forced unless required.

Of course, periodical parity checks would still be recommended, but unplanned ones will be reduced.

 

Thanks for listening.

 

My top two:

 

1. Dual parity, P+Q parity, whatever

 

2. Reliable GUI (usable with plugins)

 

 

1. Integrated encryption support

 

Filesystem- or device-based encryption layer, that will live somewhere below the unRAID share features layer. It could conceivably be under the unRAID md, or, better off IMHO, on top of md but under the disk and user shares. The most natural candidate for this is dm-crypt and LUKS. Ideally, I'll be able to configure it from the GUI, and it will take a passkey, a keyfile or both to open the container. The container could be a file (in which case the share on top of its mount point could be a user share) or a whole disk. Due to the nature of unRAID it probably can't be the whole array (which, if it were possible, would have been ideal).

 

 

Why not file based encryption? eCryptFS is easy to deploy (after some kernel tunning) and works ok.

Why not file based encryption? eCryptFS is easy to deploy (after some kernel tunning) and works ok.

 

I haven't tried eCryptFS. Did play with EncFS a little. Is there a simple way to use it in an unRAID system while still keeping the functionality of the upper layers of unRAID, i.e. the shares system etc.?

My tendency is towards a block level encryption system stashed under that layer, mainly for that reason, but this is certainly open for discussion...

My top feature requests are more near term easy to implement features.

 

1.  email support.

    Full email support is important so we can get warnings off the server to our monitoring devices.

 

2.  Better smart visibility.

    Currently there is no detailed smart visibility in the gui.

    While P+Q might help prevent a double drive failure. I think that's still a long ways away.

    The short term requires pending, un correctable and reallocated sectors to be visible

    and monitored in the gui.

    At least then, a potential drive issue becomes visible.

    If drive monitoring via SMART and email support were available, we could see the health

    of our drives and make appropriate decisions instead of depending on P+Q.

    I think P+Q provides a false sense of security when in reality a real backup is the answer.

+1 to Weebo's comment on "full" e-mail support and smart monitoring/visibility.  These actually seem a lot more feasible to implement than dual parity (which I would also like to see at some point in the future).

For me email support would be the last thing I would want.  I don't look at my email account for weeks now so it wouldn't help me.  I don't travel so don't need to check in and waiting until the end of the day to learn about a problem is better then worrying about it all day at work because I can't leave.  I also have no need to encrypt anything beyond a file by file basis and can be done at my Windows desktops.  That could change but not currently.

 

I would want the P+Q parity although I can wait for that.  I need all the space I can get anyway and this would burn up one more drive slot in the case.  It would be for a good purpose but limits the capacity max of the case by one more drive.

 

The biggest desire is to go to 64-bit which sounds like it is in the works.  But that is only really useful on my N54L where I run unRAID on bare metal and VirtualBox.  If I could get rid of ESXi and make unRAID my base OS and install a virtualization plugin that allows pass through of PCI/PCIe tuners and HDD controllers for a Windows VM then I could use it everywhere.

 

For me email support would be the last thing I would want. 

 

That's exactly how I feel about P+Q parity.

 

I use my server very heavily and it would probably slow me down even more.

The latest change Tom has implemented sounded very promising since it could increase write throughput.

Although I have not tested it yet.

 

64bit is already in the works, Tom has hinted about this a while back.

 

What I find to be of great need is a way to determine the health of your server's drives and provide alerts or notifications on it's health.  Be it the main console or some kind of remote method.

That's exactly how I feel about P+Q parity.

 

I use my server very heavily and it would probably slow me down even more.

The latest change Tom has implemented sounded very promising since it could increase write throughput.

Although I have not tested it yet.

 

64bit is already in the works, Tom has hinted about this a while back.

 

What I find to be of great need is a way to determine the health of your server's drives and provide alerts or notifications on it's health.  Be it the main console or some kind of remote method.

 

I'm with WeeboTech on this one. IMO, email (or any other notification system for that matter) and better smart visibility are very much in need for unRAID. Perhaps even preemptive alerts when smart attributes change, like current pending, reallocated, etc.

That's exactly how I feel about P+Q parity.

 

I use my server very heavily and it would probably slow me down even more.

The latest change Tom has implemented sounded very promising since it could increase write throughput.

Although I have not tested it yet.

 

64bit is already in the works, Tom has hinted about this a while back.

 

What I find to be of great need is a way to determine the health of your server's drives and provide alerts or notifications on it's health.  Be it the main console or some kind of remote method.

 

I'm with WeeboTech on this one. IMO, email (or any other notification system for that matter) and better smart visibility are very much in need for unRAID. Perhaps even preemptive alerts when smart attributes change, like current pending, reallocated, etc.

I do like the smart notification so I think I would put that ahead of P+Q for me now.  Email I already said why it isn't a priority for me.

+1 on SMART. It's all very well scheduling parity checks, but when parity checks pass and you don't check SMART you end up with this, which is what I saw the other day and I saw too just a little too late as the warranty of the drive just expired. Annoying:

 

 

That drive is really on the way out. I'm copying the data from it and will then preclear it and probably trash it.

 

My other request would be a Recycle Bin/Trash. One that works like literally every other NAS out there. Delete something, it goes in the Bin/Trash (Universal or per disk) and then you can empty later.

I know there are hacks that have been posted here, but they don't work how I expect them to. Just make a Recycler that works like every other OS in the world, please. It's just too easy to hit "Delete" on the keyboard and it's bye-bye data forever. Running arcane commands to get the data back (maybe, if it hasn't been overwritten) is not good enough.

FWIW, I had a conversation with Tom about exposing the SMART values, at the very least, as a text file on in /var/log with the serial as the filename.  While this does not solve the issues at hand. It gets us one step closer to looking at these values programatically. While I could do this myself, I felt it would be competing with access by emhttp, so I requested that Tom dump the streamed values he uses to a file (for now). Once that is available, we can write PHP scripts or modify the main page to expose these values as we see fit.

Or use something like monit to inspect them. 

 

It will still depend on getting these values OFF unRAID out to some monitored human interface.

i.e. email, pushover, growl or some other method with curl.

 

We would still need some form of agreed upon email interface. In any case. We will be one step closer once this can be done. 

 

I'll have to start another thread on SMART monitoring once this is available easily.

As I mentioned, I can do this myself, but then there is the while spin up/spin down competition.

emhttp already grabs smart information but only checks temperature for now.

My other request would be a Recycle Bin/Trash. One that works like literally every other NAS out there. Delete something, it goes in the Bin/Trash (Universal or per disk) and then you can empty later.

I know there are hacks that have been posted here, but they don't work how I expect them to. Just make a Recycler that works like every other OS in the world, please. It's just too easy to hit "Delete" on the keyboard and it's bye-bye data forever. Running arcane commands to get the data back (maybe, if it hasn't been overwritten) is not good enough.

 

Influencer updated the VFS recycle bin to a plugin which works really well

you will get a recycle bin in the settings page

which you can set to auto delete every 7 days ... or never or whatever

and you can click the icon to delete it manually

 

see here -> http://lime-technology.com/forum/index.php?topic=5446.msg269958#msg269958

 

I think it is better if we don't discuss the pro's and con's of each topic in this thread.

It would be more useful just to gather the features in this thread.

 

When proposing something you can express your opinion but others should not comment on it in this thread.

Else this thread is going to end up totally OT (with respect to the OP).

 

The evaluation of the feature list should be done somewhere else via a matrix that considers different aspects e.g. ease of implementation, expected influence on performance, timeline, etc.

 

So please let's just collect your 2 most wanted features in this thread that should be part of unRAID (not plugins).

2 ct.

I think it is better if we don't discuss the pro's and con's of each topic in this thread.

It would be more useful just to gather the features in this thread.

 

When proposing something you can express your opinion but others should not comment on it in this thread.

Else this thread is going to end up totally OT (with respect to the OP).

 

The evaluation of the feature list should be done somewhere else via a matrix that considers different aspects e.g. ease of implementation, expected influence on performance, timeline, etc.

 

So please let's just collect your 2 most wanted features in this thread that should be part of unRAID (not plugins).

2 ct.

 

+1. Thank you for this post.

Two votes for the same thing:

Updated documentation.

 

The old 4.7 wiki, etc., are in need of updates for all the changes in 5.0.

Someone elsewhere proposed that the forums be split, and a new 'branch' created for the post 5.0 discussions, with the 4.7 discussions 'archived'.

I thought that was a great idea, but didn't jump on it at the time, and now I can't find it.

 

As a noob around here, I spent hours and hours just trying to figure out which plug-ins work and which plug-ins don't work in v.5.

 

And as an old hand, you should consider how much time you spend on the forums helping noobs like me. Especially when the answer is either:

a. 'Simple Features is not compatible with release 5.0 Final.

b. update to the release version.

my two have to be:-

 

1.  email support.

Full email support is important so we can get warnings off the server to our monitoring devices. - including smart report alerting (pending and re-allocated sectors) , read errors, write errors, parity check errors.

 

2. reliable webui.

to be honest emhttp sucks badly, it crashes far too easily and can also be killed due to low memory usage, in my opinion this should NEVER happen, if your selling unraid with the idea that its a simple and reliable nas to use then the last thing you want a newbie non technical linux user to be confronted with is a dead webui!.

 

ok i know its two wishes but my third has to be a solid plugin framework, at the moment its half baked at best, i think the popularity of vlirtualising unraid tells you something, plugins just aren't stable enough to run on a "production" system, plugin(s) should NEVER stop unraid from starting up/shutting down.

 

just my thoughts.

 

Top 2 Wishlist

 

1: Double Parity

 

2: Updated/Stable Webgui with solid plugin interaction.

- Smart

- Email notification

1. Better WebUI. Stats with good looking graphs about network, array and HW usage, both avg's and realtime. For a server product, this is IMO a must. The primary function of unRAID is fileserving and redundancy, and that's something I should be able to monitor very easily without any fiddling around nor linux knowledge.

 

2. Really just more overall polish. As much as I´m impressed with unRAID, I really feel like I bought an unfinished product. Not becouse it doesn't perform the way it should, but I feel like there are to many loose ends, to many small issues like keeping the roadmap and Wiki up to date, that should be a priority for a premium product.

 

I would have put email alerts in #1 but I feel like the plugins available do a good enough job so far, but it really is something that should've been there from the start.

              Another vote for encryption, even would pay extra for have it natively implemented.

This is the most critical feature I'm waiting for, just don't want anybody to get my data in plane access even if they stole just one or more HDD's from my server.  The dual parity is not that easy to implement, so I'm willing to take some risk until it will be ready. The risk is not that great - in more then 4 years I got several HDD's fails and no data loss, thanks to Unraid!

I'll chime in here with:

 

1. increased surveillance of drive health using SMART data.

 

and,

 

2. get rid of ReiserFS, in favor a of a well supported, modern FS. I have no idea what this would take, but I have a valid reason for wanting this. One very appealing feature of UnRAID is that the data is intact and (theoretically) accessible when the drives are removed from the array. <rant> However on a Mac, which I am, there is NO WAY to access ReiserFS, no drivers, nothing. Sure I can build a Linux box (money, time, effort) or maybe rig up a virtualized Linux system (money, time, knowledge) but wouldn't it make sense to have the data accessible from the get go? I tried installing a Linux distro (Mint) to my Mac at home, and ended up with a boot nightmare from which extricating my self involved wiping the entire computer and rebuilding everything from scratch. </rant>

I think feature request ought to first and foremost be things that cannot be done by others.  E-mail and smart stuff can be crafted by people as plugins.

 

1)  Diagonal parity is #1 for me.

 

2)  64 Bit OS is #2

^^ +1

I think feature request ought to first and foremost be things that cannot be done by others.  E-mail and smart stuff can be crafted by people as plugins.

 

1)  Diagonal parity is #1 for me. ?

 

2)  64 Bit OS is #2

 

is this a different type of P+Q parity ?