Of course for any features you want you must pay…modules…spam filters…etc…the question is not how the API is attractive or not for a developer, it is just a question of money…there are so few people who do things for nothing.
I see ClearOS 7 Community for $0/month.
One thing I’ve always disliked about them, they make it really time consuming to figure out what’s what.
I have to say though, addressing the original question.
Wouldn’t migrating away from esmith, be a multiyear affair? I mean, jeez, not sure that’s what I would want to invest in, seems easier to invest the time to learn esmith… for me, who needs to solves client issues and desires.
The ClearOS community edition is still existing and there are a lot of module that are free… lot of more than Untangle for example.
But apparently there will be a change with the ClearOS 7 and the home edition.
We will see if users will come from ClearOS in the near future…
It’s the time to improve the tutorials to make the e-smith system more afordable 
1 Like
uh oh
Users of Community edition will have a 30-day window to upgrade to Home or Business version.
Is the e-smith layer can deal with the upcoming Docker?
I think it’s possibile. Moreover Docker can fit Shorewall too!
continuing this topic… someone seems not to realize that “the magic” is done by the business logic inside e-smith layer but by the UI…
one time more: stop thinking about the UI…
let’s think about installing a new service on NS… not one available from NS’ repos, just a rpm
it’s just a question of
yum --enablerepo=extrepo install myservice
once done, we have myservice installed…
the first things to do are:
- enable it at startup
- configure firewall to let the service be available from van on port tcp 1234
using the classical approach, we need to poke with chkconfig and iptables rules
using e-smith layer, it’s just a matter of
db set mysservice service status enabled TCPPort 1234
signal-event runlevel-adjust
signal-event firewall-adjust
done… no need to know how the underlying system works (sysvinit, systemd, whatever), nor hot to deal with iptables syntax…
now, we have likely to configure myservice changing some parameters into its conf file…
let’s say that one of the requested parameter is the domain name and/or the ip address
no rpoblem, we edit the conf file, restart service… all is working as expected… fast forward, an year later, we need to change the ip address… a trivial task, done in 30 secs via web GUI…
oh, dear, wait… myservice doesn’t work anymore… wtf? damn, forgot to change its config…
via the templating system we can avoid to be in such a situation… ip address is just a variable, after changing it everything still works flawlessy.
usually, all the fragments, db values and so on are stored into a nethserver-myserver package… stephdl just showed us what such a package is…
the templating system has some advantages more…
let’s say I install nethserver-mynewservice… I configure it via webgui… so far, so good… one day I need it to be configured in a different way… with templating approach, and the custom templates/fragments, I can act either in adding custom parameters (for example some one not available via web GUI) or overriding the logic of the original template/conf file… and all is done via small snippets of code (they can be just plain text in some cases, no need to use any kind of language) that are added automatically to my backup…
with such an approach, user can customize at 100% his server, being sure his editing won’t be lost on a server restore, and all without (I repeat) dealing with conf file syntax or location, without knowing any kind of cli commands more than calling an editor, without worrying about the underlying O.S.
tomorrow, moving to centos7 will be “just” a matter of adapting the logic to the different subsystem, but all the e-smith layer exposed commands will be there, working as expected…
in my personal experience, using this approach, I migrated a SME server 6.01 (based on RH7.3) setup in january 2006 to a SME server 7.1 (based on centos4) to a SME8 (centos5) to a SME9 64 bit (centos 6.6 64) (and with 5 hw changes…) with a global downtime (in almost 10 years) of… maybe 10 hours… and I never had a problem with new conf files and/or daemon changes…
NS is based on the same logic; it needs some time to be undestood, but once you’ve got it, it’s more than rock solid
5 Likes
I think the someone it’s me 
I`d learned Linux and before Unix, I’m still looking for BSD.
From Linux, I’d seen Debian, Archlinux, Gentoo, Slackware and now CentOS.
And yes, when I’d installed Nethserver, it was for pure commodity for not to install a minimal CentOS and a webmin interface and configure all by myself… To gain time.
I like a lot the KISS approach of Slackware, so an e-smith layer, that seem a complex bunch of template with a database… look unnecessary.
You give me a real good explanation, thank you.
But, independent of the quality of the e-smith layer, why there’s so few developers?
In my IT live, I’d seen lot of good projects that didn’t go ahead cause of the lack of developer…
Could you share some example?
Borland Geoworks https://en.wikipedia.org/wiki/GEOS_(16-bit_operating_system)
IBM Os/2 Warp https://en.wikipedia.org/wiki/OS/2
They seems quite old and closed source projects… 
I was young
2 Likes
first of all, all the layer is in perl, the web GUI too… this is both a strength and a weakness… the first because compared to php (for example), perl IS a real programming language, while php is a bunch of (insecure) instructions 
the second because there are zillions of php code monkeys and many less perl coders…
another reason is that SME is 100% community driven, meaning that all the money come from donations… no one is payed for his work…
finally, and this is an issue with all OOSS, in every community the 99% of members are just users or, in the worst case, just beggers… no one is interested in being involved… most of users come to the forum, ask for help and just disappear once they’re done.
here the numbers are still relatively small, but on large scale that % is real an quite optimistic
Are we missing some security patches from upstream? 
no… it’s not a matter of patch but of design…
php is widely used but poorly designed and unsecure
To continue the discussion.
If the CentOS base is not a problem
If the Webgui is not a problem…
And if this e-smith layer is really good…
Why Nethserver is not wider adopted?
Is the all in one concept is dead at the virtual machine’s era?
In this case why ClearOS has success?
Is a problem of strategy? or comunication?
I see nothing about the e-smith layer in the main site perhaps we need to adapt the comunication, but at the same time there’s now reviews about Nethserver on the net and in newspapers.
Or is Nethserver or the e-smith layer is for a niche ( as a niche market) ?
You’re right, I ask myself the same question every single day.
And yours are perfectly fair questions… But I have no answer. Maybe the product is too young? Remind you that this community was built from scratch 6 months ago and the last month we have counted 1,7k posts, 150 active members and about 500 registered. Plus 16.500 download just for 6.6.
We’re still small but we could grow up togheter with a more and more active community, a better product that is steadly improving, more content/blog post/reviews and maybe a killer application.
I give you an example:
Docker is one of the most popular software of the last two years… Is it easy to develop? Nope. Is it easy to use? Yes. And it’s famous because it’s smart, it saves costs hardware/power and brings several new things to the table that the earlier technologies didn’t
SO, we don’t need to make NethServer good, we need to make NethServer AMAZING
Once @Ctek said me “guys that’s a great product why I found it just today? Why Zentyal and Clearos are so famous compared to NethServer?” He was right!
That’s said, things above are the keys IMHO.
And togheter we can do it, that’s my belief.
4 Likes
Jim…
seems to me you are trolling a bit 
NS is still young, so you can’t expect that it is well know and widely adopted.
then… we’d discuss about the “adopted” meaning… an home user can adopt it and leave it when and how many times he wants…
but NS is mainly an enterprise aimed distro… in that perspective, no one will adopt it without trying it, knowing how it works, how it’s developed… and you can’t do it in 3 days…
nowadays there are dozen of user friendly linux distro you can install in 3 clicks, without even answering a question… that’s good…
but the real thing comes when something stops work or work in the way you don’t like…
would you install an unknown distro (unknown -> you don’t know how it works internally) to a customer that manage 12 tb of data? or one that send and receive 5000/10000 mails daily?
what if it stops working?
don’t get me wrong, but you’d adopt a different point of view, more business continuity oriented