Adopting a different OMV approach

Hi,

Can just agreed with you.

First of all, these are some reflections and thinking that does not criticism the wonderfull job already done by Volker and Marcel. OMV is a great software and I wish it can have the longuest life it can have. I worked with Marcel, at my level, to develop and test the omv-plugins.org ones, done pratically all the French translations of OMV and all the plugins, and I know how it can be some time consuming, almost if you have one family to live with
But here are, for me (and probably my needs), the way to go for OMV to be a REAL NAS solution as it can/should be in the near future...

With some "killer" core and/or plugins that are missing today, OMV can easily be a first class NAS. Unfortunatly, as you said, it's a little difficult at the moment to write some, even if you're used to be a java dev like me !

Personnaly, I think it missed a real internal Web File browser, to let people that can't have a FTP/CIFS/SMB access to their NAS using it through the internet (e.g. from works and so on...). The idea here is to put something like eXtplorer in the core of OMV. The code of eXtplorer is already PHP and just need to be updated to take OMV user rights system. Volker, if you read me, please, take a moment to think at it
We need a tree view to navigate, a files window to make some actions like coping/pasting, drag & droping files between dirs, and the ability to upload/download a set of file with zip compression through HTTPS over the internet.
For me, this MUST be in OMV core, and not delegated to a plugin as it's a necessary app for any NAS nowadays.

Secondly, a good DLNA server solution would help getting all the home users. Today, allmost all NAS are build to deliver Musics/Photos/Vidéos, so the DLNA server is the center application.
I think that MiniDLNA is a good start and does the job simply (essential), but IMO, needs to be develop to be a killer Home server app. I was thinking of moving the database background of Minidlna to mySQL (manage to do it with the structure and datas from sqlite very easily) to let other app search/modify the metadatas of medias (internet search and completing, etc...), but unfortunatly my C knowledge left me 15 years ago
Anyway, once more here, the key thing is to find a C developper to extend/modify the source code of miniDLNA to let it be more database independant. I looked at it and clearly see what needs to be done (by comparing it to Mediatomb code which is really database independant), and it's not very complicated, but clearly don't have the C knowledge

At the end, a good backup solution can be added by a CrashPlan or BackupPC plugin, to backup your home computers.

With that, you will have a solid ready to go NAS for 90% of the home users... but OMV needs more devs, clearly!

Zitat von "XoN"

From my user perspective, I think that in order to grow, you need more features… more plugins, such as owncloud, sickbeard + sabnzbd + couch potato, crash plan…. those kind of features to make OMV more "famous" or "known" (I don't know how to say it but I thing you get the point :D)… At the same time, this would bring more developers to help you improving the system.

Sorry to barge in on this very important topic, being new to the forum and all, but I have to share some thoughts.

As a start-up project or business, consider the extremely limited resources you have (money, time, experience, people) and the limitations of the choices you can make to use these resources. Most start-ups I have seen have failed miserably because they were externally focused, trying to have either a better product, a better price, or both vs. another mature product. It is impossible to compete directly with the bigger companies or projects already established in the same industry or service; they already have a well developed base to build on, more resources, and they are a lot more cost-effective.

It is tempting, but it's very, very wrong to not define some initial boundaries for what you want to do. Don't spread your resources thin by trying to incorporate many additional features. Most of them are not mission-critical, but just "nice to have". On the other hand, the constraint of available resources to create all these features means the product will not excel at any of them. And that, kind sir, is a lot of wasted energy into a product which fails to meet basic requirements of ANY user.

The start-up should be internally focused: define what you stand for and make the best out of it with what you have. Do just a few things at first, and do them well. Be recognized for creating a specialized tool which is THE most efficient and powerful at that particular thing. Grow a community of fans and supporters and outsource everything which is not critical for the product core, allowing others to bring in enhancements through plug-ins. When you have a solid core that will support your growth, start expanding to new features - maybe incorporate some of the additional functionality developed and matured by others into your core.

I got stuck with OMV for 4 simple reasons:
1) Cost efficiency. Why buy an expensive NAS with mediocre performance or a VERY expensive NAS with decent performance, when I already have some available hardware and just need the software to make it work?
2) Speed. OMV is by far the highest performance NAS software I have tried.
3) Ease of use. I'm not afraid of a command prompt, but if it can be done a lot faster with a few clicks in a good web UI, why not?
4) Reliability. I want the piece of mind that a crash will not make my data disappear. I've read enough horror stories from various dedicated NAS machines (even Synology) which lost data partitions or couldn't rebuild the RAID or it took a few days to rebuild a replaced hard drive, or from FreeNAS whose flaky reliability practically requires you to have a full backup of everything on the NAS.

Universal access? Sure, that would be nice... But then, 99.95% of the time I use the NAS from the local network. When I travel, I either carry the necessary data with me on a stick, or store it in the cloud, or even access my home computer remotely. And for that 0.05% of the time when I might actually need a file off the NAS, I can securely FTP in on a high non-standard port. And that's just an example.

It's a no-brainer that I would definitely like more features But they will come in time, as the product develops. I've seen this course of action with SmoothWall, a dedicated software router solution.

What the community CAN do to help is to spread the word. Use OMV, show its strengths, showcase it on blogs, create some buzz. People WILL come.

Thanks for reading Greetings and all the best to the OMV team!

Zitat von "puterfixer"

Sorry to barge in on this very important topic, being new to the forum and all, but I have to share some thoughts.
.../...
What the community CAN do to help is to spread the word. Use OMV, show its strengths, showcase it on blogs, create some buzz. People WILL come.

+1 I fully share your point.

Offering more features based on plugg-ins... why not if time to develop, test and deploy it is not taken on time required to ensure that OMV as "storage platform" is as reliable and efficient as it can be
Doing more than what is tightly linked to storage itself is very risky because you will always find here and there different views about what could/should be the best solution for "cloud" (I really don't like this word ) or web interface for file management, or webmin like approach ... while there is still room for improvement looking only at storage features like SAN, NAS, redundancy, backup...

I do understand that without new features from time to time, this is difficult to attract new users but reliability is really the first goal, then efficiency (performance) then administration easiness.
As puterfixer wrote, IMHO other features will follow later.

Hi all,

It's been a long time since I went here. I now visit from time to time where I was visiting each day in the last few month.
Just add 2 cts here because when I now read the forum, I can only see people having problem because of plugin non ported, and so long...

Believe me, I love OMV, but it simply missed some "core" feature to be a real alternative to other NAS system manufacturer.
I'm a lead system & network computer person since 20 years, and clearly, love spending my time to find solutions to computers problems, but most of people just want their NAS to work simply, without allways tuning or tweaking it.
When you spend 1 hour to configure minidlna to have a correct DLNA solution, or spend hours to manage redondant storage/backup, torrents, etc... (put there what you want), we're clearly missing the target.

I know that developping plug-in is not an easy way to go with OMV, and IMO, it's probably here that the problem is. Marcel did (do?) a great work, but clearly we missed some devs at this point to secure/update the plugins that already exists. I know it's time consuming (I use to test/validate the greyhole plugin and believe me, it takes time), but something needs to be done if you don't want OMV to be more than a PHP GUI to debian.

If OMV is said to be entreprise targeting, ok, you should have a person to manage it, tune it...
If OMV is said to be home targeting, clearly, most of us prefer spending their time to other things... (most )

Well, I still continue to follow and support OMV, don't get me wrong (Volker, love your work, believe me), but I cannot help thinking that we are missing something here.

We are missing more Ian Moore's. I think one of the best things that could be done would be some documentation on making a plugin. A developer should not have to spend an inordinate amount of time in order to add their project into omv. Also, perhaps we as a group are failing in reaching out to other developers to ask for help. Has anyone contacted a developer of the program that they would like to see added as a plugin? Volker is not the problem. I think it is the community that needs to do more at this point. In all the social media today we can do more there alone to promote OMV.

Zitat von "tekkbebe"

Has anyone contacted a developer of the program that they would like to see added as a plugin? Volker is not the problem. I think it is the community that needs to do more at this point. In all the social media today we can do more there alone to bring more attention to OMV.

I can only speak for myself and a little bit for Marcel as we were developing the Greyhole pluging together.
When we ask Gauillaume BOUDREAU some help for Greyhole, he just answers quick and sharp and that was very helpful. He also added some features we needed! But he doesn't want to spend time on the plugin, as he surely have to fix/update some stuff on the Greyhole core, or what else he wanted to do.
That's a point of view I can clearly understand, and that I think most of the "core" apps dev will go.

Second exemple, I was planing to improve the minidlna plugin to be more integrated with OMV. Ask some help on miniDLNA forums to devs, but nver get a single response! That the way life goes...

So we have to do it ourself, the best way we can, but clearly, plugin aren't easy to write and update...

I'll second and third the plugin dev comments. Right now I'd like to start building a plugin, but here is the situation as I see it from doing lots of searching etc.

Plugin architecture is not documented anywhere I can find. Yes I can reverse enginner the deb packaging and make something work. This is not the same as knowing where things should go or coding/deployment standards for a plugin. There is some documentation on coding standards, 4 spaces etc, but nothing on OMV architecture, web standards etc. If my plugin needs to store config, what is the preferred way to handle it? Is there an expected segregation of configs, such as program specific things being kept in /etc/someconf, but OMV plugin specific items being stored ... I don't know where

Now even if I take that apart, seeing as an enhanced and easier to use plug-in api is slated for 0.5 - that makes an attractive target for a new developer. However I can't find any doc on how to get a preview or build my own 0.5 buiid. If I wanted to go ahead and build one against 0.4.x well I can only find the 0.5 api guide. This is probably just me, but when I want to take on a new project, I usually try and target the next version, and there seem to be compelling reasons to do just that with OMV.

I don't mean to criticize anyone, I've gotten some helpful responses on the forums etc, but the barrier to entry on OMV plugin development is very high. Right now the path I see is to reverse-engineer a plugin for 0.4.24, learn an API that will be deprecated and possibly create a plugin that will have a very short life span depending on when 0.5 comes out. If there is anyway to get a little hand holding with the initial plugin dev, I'm very happy to create a plugin dev guide along the way, I'm even happy to own and maintain it for say the next two version of OMV so that the investment in dev time is well used. I'd really like to see a larger community of devs get involved with OMV, myself included. If that means volunteering to do documentation I'm all for it.

Maybe I'm just lazy, but I at least try to be constructively lazy. I have already built out a dev environment with git and PDT, I've setup a VM based test bed and now have 3 OMV's running on my network, I've grabbed the source for 3 plugins and spent some time reading through the structure, etc. Hell, I'm about to buy another n54l just so I can have a dedicated bare-metal test box.

Zitat von "votdev"

At this point it makes sense to start wrining docs how to develop plugins.

Hear, Hear!!!

Zitat von "votdev"

..... Maybe in future we can change the focus, but at the moment i will not do this step. I knew that plugins will help to spread OMV, but it will bind resources too much if you try to do too much at the same time. Because of this the plugins shipped with OMV are focused on home scenarious...

Thank you for replying, and I don't want to assume anything, but that reads to me like open plugin development is not really a goal and may even not be desirable to you. If that's the case, that's entirely fine, I will happily leach code, hack up debian and provide testing Considering I'm not the only person on these boards I've seen make offers of help on plugins, clarification on this would be appreciated (but not necessary, hell you've given me a great little NAS setup)

I think Ryeco's on point. Volker is a busy man. Remember this is a side project for him and he has a job. Er ist nur ein Mensch.

Yeah i could have read it that way too, but it brings me to my 2 rules of community forums.
1) Always assume the reply was intended to help
2) When in doubt, get clarification politely.

I'm new here, so i don't have context to help me interpret Volker's statement. I completely understand it's a side project, heck I don't even code for a living anymore so even maintaining multiple dev environments is pure hobby (I also do some android dev). That's why I only asked for clarification, not please do it now I also understand that adding a new resource to a software project can temporarily slow down dev, but I am making a very sincere offer of real man hours. (I make no warranty as to those being useful man hours). Honestly, all I'm trying to figure out is the least disruptive way to contribute some time and some code. Forgive my OCD, it's just when I get a bug... I really really get it. I'll chill from here on out.

BTW: I also tore apart the miniDLNA plugin yesterday as Ryecoaaron suggested. Unfortunately I'm missing something, most of it makes sense, but the github repo can't be everything - it seems like there are other pieces not in the git repo, nor do the dependencies seem defined in the package. It's also entirely possible I'm making very bad assumptions as to what is in miniDLNA - since I haven't used it. When I revist plugins, I'd probably tear apart the transmission one, since I've had a decent amount experience with that both on pc's and embedded systems.

This is a great project and I think the name is more of an issue. I discussed this with Volker. Let's face it, "free" in a name makes most people's mouths water like Pavlov's dogs. I believe in Volker and in time we will overcome the name. We hear you and encourage your help. Volker is having some issues on .5. Once these are worked out I'm sure he will take more time on the documentation of creating a plugin. For now everyone just keep channels open and share info. on forum. I know Ryeco is workin on a small plugin now and he can guide into findings things.

The minidlan plugin, openmeidavault-minidlna, does not include the minidlna package. There is another package for minidlna in the repo.

Zitat von "ryecoaaron"

The minidlna plugin also installs the minidlna package for the binary and init script.

I'm either blind or an idiot or just had one too many windows open in eclipse- or not enough coffee yesterday or any of a million excuses, becase it's right freakin there in control My bad.

Might I make a small suggestion? Can we create a "dev" forum? Even if 0.5 and plugin doc isn't ready to go, this just seems like it would be good to have a dedicated forum for these discussions. I'd rather not show off my blatant inability to read a depends line in "General Discussion."