Discussion:
NethServer: something new to play with
(too old to reply)
Filippo Carletti
2013-01-09 11:24:58 UTC
Permalink
Smeserver users and developers,

during the last months, we (Nethesis) tried to build what we think
could be a good SME-oriented distro:

NethServer

Our goals were:
- simple configuration
- quick trouble shooting
- fast development of new features

We think we can achieve those goals with:
- a new dynamic, streamlined user-interface
- standard components, tested widespread configurations
- minimal differences with upstream

We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php. We rebuilt the mail server using
postfix, amavis and dovecot.

We are still working on many features and writing documentation for
contribs developers, but we hope to have reached a point were an alpha
release is good enough to have your opinion.
This alpha release is well tested but not ready for production environments.

NethServer is released under the GPL, development is open, developers
are welcome.
You could find more informations on the website:

http://nethserver.nethesis.it

Please use the NethServer group for support requests, comments and questions:

https://groups.google.com/d/forum/nethserver

Thank you.


--
Ciao,
Filippo
Hsing-Foo Wang
2013-01-09 13:59:37 UTC
Permalink
Ciao Filippo!

very interesting. I've downloaded the iso and installed it in Parallels
virtual machine. Installed and booted fine. Congrats on this effort.

Questions that will pop-up:
1. Is it a fork of SME Server
2. Who is the sponsor and how it is organized?
3. etc.

- Keyboard used is non-us intl. Maybe it is Italian?
- Webserver (hhtpd) is down, can't reach manager
- DNS during install is set to 8.8.4.4.8.8.8.8
- Yum update has a faulty centos mirror

I'm not using Google+ and will not. Is there is another way to discuss
Nethserver?

TIA
-HF


On Wed, Jan 9, 2013 at 12:24 PM, Filippo Carletti <
Post by Filippo Carletti
Smeserver users and developers,
during the last months, we (Nethesis) tried to build what we think
NethServer
- simple configuration
- quick trouble shooting
- fast development of new features
- a new dynamic, streamlined user-interface
- standard components, tested widespread configurations
- minimal differences with upstream
We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php. We rebuilt the mail server using
postfix, amavis and dovecot.
We are still working on many features and writing documentation for
contribs developers, but we hope to have reached a point were an alpha
release is good enough to have your opinion.
This alpha release is well tested but not ready for production
environments.
NethServer is released under the GPL, development is open, developers
are welcome.
http://nethserver.nethesis.it
https://groups.google.com/d/forum/nethserver
Thank you.
--
Ciao,
Filippo
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Filippo Carletti
2013-01-09 15:06:45 UTC
Permalink
Post by Hsing-Foo Wang
very interesting. I've downloaded the iso and installed it in Parallels
virtual machine. Installed and booted fine. Congrats on this effort.
Thank you.
Post by Hsing-Foo Wang
1. Is it a fork of SME Server
Yes, but we kept only templates, events and config db. Everything else
is written from scratch,
Post by Hsing-Foo Wang
2. Who is the sponsor and how it is organized?
Nethesis.
We sell support for smeserver since 2003, we have hundreds of
customers in Italy, we have a team of trained syadmins that give email
and phone support.
We are a solid and profitable company, you can find my name (and
Federico and Giacomo) in the changelog of some packages of smeserver.
Post by Hsing-Foo Wang
- Keyboard used is non-us intl. Maybe it is Italian?
Yes. You can modify /etc/sysconfig/keyboard.
Post by Hsing-Foo Wang
- Webserver (hhtpd) is down, can't reach manager
Access the web interface using a browser at: https://<your_ip>:980
Post by Hsing-Foo Wang
- DNS during install is set to 8.8.4.4.8.8.8.8
We choose google public dns.
Post by Hsing-Foo Wang
- Yum update has a faulty centos mirror
Do you have the error?
Post by Hsing-Foo Wang
I'm not using Google+ and will not. Is there is another way to discuss
Nethserver?
A google group is a "standard" mailing list, you only need an email
address, whatever it is.
If you have a google account (not G+) life is simpler, but you can
send an empy email to:
nethserver+***@googlegroups.com


--
Ciao,
Filippo
Hsing-Foo Wang
2013-01-09 17:34:47 UTC
Permalink
My mistake, due to removing the google DNS. It looked funny to me. So
'green' could not resolve.
Post by Hsing-Foo Wang
- Yum update has a faulty centos mirror
Do you have the error?
Charlie Brady
2013-01-09 15:55:08 UTC
Permalink
Post by Filippo Carletti
We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php.
I think that choice was very unfortunate. I applaud your initiative, but I
wish I had had an opportunity to provide some guidance on the technology
of your UI replacement. Please have a look at:

http://mojolicio.us/
http://mojocasts.com/e5

and read:

http://me.veekun.com/blog/2012/04/09/php-a-fractal-of-bad-design/

Whatever you are doing in PHP can be done more securely and more reliably
in perl, python or ruby. I hope that it is not too late for you to
reconsider.

Best wishes, and congratulations.

---
Charlie
Hsing-Foo Wang
2013-01-09 23:37:58 UTC
Permalink
I kind of agree with Charlie. The 'kind of' comes from the fact that 'we'
have a rock solid perl skill level embedded within the SME Server community
(devs) and the philosophy about security and simplicity proved itself since
the very very early days, based on the use of Perl. (Not many distro's can
say the same...)

I recently submitted a patch. I am not a programmer, but along the way I
'slightly' started to 'understand' the perl language, it's consistency and
power. Hence I consider my patch as a ver early beta and poor attempt.
Nevertheless, while trying to achieve my goals I considered many things,
even wrote a whole new bash ''console app', when I realised I needed to
re-use the excellent code already there and pushed myself to look at the
existing possibilities. So I went dow from 300 lines of code to merely 10
lines of code for this specific patch.

I also remember the days, way back around 2005 when Gordon R. was still
around, we briefly discuss specific distro 'SME Server upgrade', meaning
when you install Centos, 'yum install sme-server'. It seems that after
almost 10 years you have done this :-)

So I second the request of Charlie if it would be possible for you/your
team to reconsider using PHP AND if you guys would join forces with the SME
dev team, and specifically the other way around...

If so, Nethserver could possibly SME Server-NG

Again compliments on the achievement, a welcomed innovative development!!

Ciao,
-HF




On Wed, Jan 9, 2013 at 4:55 PM, Charlie Brady <
Post by Charlie Brady
Post by Filippo Carletti
We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php.
I think that choice was very unfortunate. I applaud your initiative, but I
wish I had had an opportunity to provide some guidance on the technology
http://mojolicio.us/
http://mojocasts.com/e5
http://me.veekun.com/blog/2012/04/09/php-a-fractal-of-bad-design/
Whatever you are doing in PHP can be done more securely and more reliably
in perl, python or ruby. I hope that it is not too late for you to
reconsider.
Best wishes, and congratulations.
---
Charlie
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Charlie Brady
2013-01-10 00:10:39 UTC
Permalink
Post by Hsing-Foo Wang
So I second the request of Charlie if it would be possible for you/your
team to reconsider using PHP AND if you guys would join forces with the SME
dev team, and specifically the other way around...
Please everybody, go and have a look at Mojolicious. It's what I strongly
recommend for new UI developments. Very, very well maintained and tested.

http://en.wikipedia.org/wiki/Mojolicious

http://www.slideshare.net/sharifulin/mojolicious-the-web-in-a-box

https://github.com/kraih/mojo/wiki/Projects-and-Companies-Using-Mojolicious
Charlie Brady
2013-01-10 02:41:45 UTC
Permalink
Post by Charlie Brady
Post by Hsing-Foo Wang
So I second the request of Charlie if it would be possible for you/your
team to reconsider using PHP AND if you guys would join forces with the SME
dev team, and specifically the other way around...
Please everybody, go and have a look at Mojolicious. It's what I strongly
recommend for new UI developments. Very, very well maintained and tested.
http://en.wikipedia.org/wiki/Mojolicious
And especially here:

http://mojolicio.us/perldoc

I see that there are 32000 lines of php under /usr/share/nethesis, so
assume that Filippo will be reluctant to change tack at this stage. But I
look at https://servername:980/ and what I see is neither very elaborate
nor complete. So perhaps we need to prototype something in
perl/Mojolicious::Lite to show that it can be done efficiently and
effectively.

---
Charlie
chris burnat
2013-01-10 04:17:28 UTC
Permalink
Post by Charlie Brady
Post by Charlie Brady
Post by Hsing-Foo Wang
So I second the request of Charlie if it would be possible for you/your
team to reconsider using PHP AND if you guys would join forces with the SME
dev team, and specifically the other way around...
Please everybody, go and have a look at Mojolicious. It's what I strongly
recommend for new UI developments. Very, very well maintained and tested.
http://en.wikipedia.org/wiki/Mojolicious
http://mojolicio.us/perldoc
I see that there are 32000 lines of php under /usr/share/nethesis, so
assume that Filippo will be reluctant to change tack at this stage. But I
look at https://servername:980/ and what I see is neither very elaborate
nor complete. So perhaps we need to prototype something in
perl/Mojolicious::Lite to show that it can be done efficiently and
effectively.
---
Charlie
My congratulations and thanks to Filippo Carletti and his crew for a
nice New Year present to the Open Source Community.

Charlie, should you be willing to undertake this prototyping and more
generally bring your invaluable expertise to this project, I would be
very happy helping with testing and what else I can do to the best of my
abilities. This does not mean I am leaving the good ship SME just yet,
I believe we have an obligation to our users. A simple and effective
transition path will be required, no to forget maintenance of version 8
until migration to a new platform is implemented. Another topic;-)
Hsing-Foo Wang
2013-01-10 06:57:56 UTC
Permalink
My congratulations and thanks to Filippo Carletti and his crew for a nice
New Year present to the Open Source Community.
Indeed!
Charlie, should you be willing to undertake this prototyping and more
generally bring your invaluable expertise to this project, I would be very
happy helping with testing and what else I can do to the best of my
abilities. This does not mean I am leaving the good ship SME just yet, I
believe we have an obligation to our users. A simple and effective
transition path will be required, no to forget maintenance of version 8
until migration to a new platform is implemented. Another topic;-)
I will try to pitch in, but unfortunately I'm not a 'perl monkey', so I may
be able to pick up another task.
Ian Wells
2013-01-12 05:47:47 UTC
Permalink
I would be happy to work on any sustainable continuation of SME Server,
which may be NethServer.

First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.


Ian
Hsing-Foo Wang
2013-01-12 19:49:51 UTC
Permalink
and that is good news Ian.

I'm sure many more readers appreciate you reaching out to the NethServer
team, as well as appreciate all the hard work you, Shad, Charlie, Chris and
others (Come to think of it, there's a great team out these) are doing.

Domo arigato gosai mas!
Post by Ian Wells
I would be happy to work on any sustainable continuation of SME Server,
which may be NethServer.
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Ian
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
chris burnat
2013-01-13 04:59:59 UTC
Permalink
The question we need to ask ourselves: is SME coming to an end,
warranting throwing away so many years of hard work from so many
distinguished developers, some of the best documentation and contribs
around, and above all a thriving community? Our community.
I think not.
chris
Post by Ian Wells
I would be happy to work on any sustainable continuation of SME Server,
which may be NethServer.
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Ian
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Graeme Robinson
2013-01-13 06:00:42 UTC
Permalink
The question we need to ask ourselves: is SME coming to an end, warranting
throwing away so many years of hard work from so many distinguished
developers, some of the best documentation and contribs around, and above
all a thriving community? Our community.
I think not.
agree, with an exclamation mark. Postfix over qmail? I don't think so.
Stefano Zamboni
2013-01-13 09:40:26 UTC
Permalink
Post by Graeme Robinson
The question we need to ask ourselves: is SME coming to an end, warranting
throwing away so many years of hard work from so many distinguished
developers, some of the best documentation and contribs around, and above
all a thriving community? Our community.
I think not.
agree, with an exclamation mark. Postfix over qmail? I don't think so.
I think we all are missing the point.. this is not a qmail vs postfix
question..

AFAIK SME development is almost stopped

CentOS5 is getting old, 6 is already out, stable..

do WE have resources to move SME to CentOS6?

should we waste time in "$our_favourite_daemon" vs "$new_daemon" war?

we have knowledge, know how.. since we are not asked to change
distribution we should be able to move to the new, aren't we?

we have an opportunity here: we have a company that support
development.. can we, really, miss this train?

I don't think so

all IMVVHO

ciao

Stefano
Gordon Rowell
2013-01-13 10:03:28 UTC
Permalink
Post by chris burnat
Post by Graeme Robinson
Post by chris burnat
The question we need to ask ourselves: is SME coming to an end,
warranting
Post by Graeme Robinson
Post by chris burnat
throwing away so many years of hard work from so many distinguished
developers, some of the best documentation and contribs around, and
above
Post by Graeme Robinson
Post by chris burnat
all a thriving community? Our community.
I think not.
agree, with an exclamation mark. Postfix over qmail? I don't think so.
I think we all are missing the point.. this is not a qmail vs postfix
question..
AFAIK SME development is almost stopped
CentOS5 is getting old, 6 is already out, stable..
do WE have resources to move SME to CentOS6?
should we waste time in "$our_favourite_daemon" vs "$new_daemon" war?
we have knowledge, know how.. since we are not asked to change
distribution we should be able to move to the new, aren't we?
we have an opportunity here: we have a company that support
development.. can we, really, miss this train?
As a long time lurker since I stopped being able to commit time as an
active developer here, I find this announcement both exciting and sad. It
is both a new beginning and a death, as a fork of SME Server (which this
is), means that one or both will die. While I applaud the move to a new
base, I'm not sure of the rationale behind replacing components that have
worked well for many years. I also don't understand why this effort wasn't
put into SME Server 8. Supporting both simply won't happen and the
guaranteed fragmentation of the community is sad.

As I am no longer able to commit time to this project, I have no say in its
direction. I wish both projects all the best, but forks are never a good
idea <http://mako.cc/writing/to_fork_or_not_to_fork.html#id2447893>.

Thanks,

Gordon
Post by chris burnat
I don't think so
all IMVVHO
ciao
Stefano
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Charlie Brady
2013-01-13 19:24:52 UTC
Permalink
Hi Gordon! Happy New Year.
Post by Gordon Rowell
As a long time lurker since I stopped being able to commit time as an
active developer here, I find this announcement both exciting and sad. It
is both a new beginning and a death, as a fork of SME Server (which this
is), means that one or both will die. While I applaud the move to a new
base, I'm not sure of the rationale behind replacing components that have
worked well for many years. I also don't understand why this effort wasn't
put into SME Server 8. Supporting both simply won't happen and the
guaranteed fragmentation of the community is sad.
Well said.

I don't know how set Filippo and colleagues are in their direction and
their technology choices. It could be that the public alpha is a
technology test platform and some of it they will be prepared to throw
away. But I doubt that is the case - they are obviously a long way down
the road of making major changes to many parts of the distribution, for
reasons we can only guess at.

---
Charlie
Filippo Carletti
2013-01-14 16:49:47 UTC
Permalink
[English is not my native language, please bear it in mind if some of
my sentences may seem incorrect/inappropriate]
I also don't understand why this effort wasn't put into SME
Server 8. Supporting both simply won't happen and the guaranteed
fragmentation of the community is sad.
Gordon, nice to hear from you.
I think that you (and many others reading here) are missing the
discussion that took place in September/October 2009 (more than three
years ago) in the "Board" mailing list.
There, we talked a lot about the future of SME and came to nothing.

During the last years, I, Federico and Giacomo contributed to sme
providing bug fixing, patches, new features, testing and a bit of
support. Sometimes, we had our patches refused, especially for new
features. It forced us to fork some sme packages, we did that with
disappointment, we tried really hard not to fork, but we have a
business and our customers always have a point, so if the feature they
were asking for was useful we developed it, submitting a description
to contribs bugzilla, providing code and building packages if code was
approved.
We have been and still are very open to discussion.

But the discussion that took place in the close Board mailing list led
us to think that a new effort to further develop sme was doomed to
failure.
We set some key points (close to upstream as much as possible, proven
technical base, well known and widely used components), we aimed at
some goals (simplicity, user and sysadmin friendliness, easy
troubleshooting) and tried to imagine where we wanted to go.

Now, we have reached a point were we have something working, and we
wanted to show our work.
It's an alpha release, it has bugs, it is subject to changes. But it's
modular, easy to customize and modify.
It's GPL, every issue is publicly discussed on redmine.

I wish everybody a happy new year.

--
Ciao,
Filippo
chris burnat
2013-01-14 20:24:03 UTC
Permalink
Post by Filippo Carletti
[English is not my native language, please bear it in mind if some of
my sentences may seem incorrect/inappropriate]
I also don't understand why this effort wasn't put into SME
Server 8. Supporting both simply won't happen and the guaranteed
fragmentation of the community is sad.
Gordon, nice to hear from you.
I think that you (and many others reading here) are missing the
discussion that took place in September/October 2009 (more than three
years ago) in the "Board" mailing list.
There, we talked a lot about the future of SME and came to nothing.
During the last years, I, Federico and Giacomo contributed to sme
providing bug fixing, patches, new features, testing and a bit of
support. Sometimes, we had our patches refused, especially for new
features. It forced us to fork some sme packages, we did that with
disappointment, we tried really hard not to fork, but we have a
business and our customers always have a point, so if the feature they
were asking for was useful we developed it, submitting a description
to contribs bugzilla, providing code and building packages if code was
approved.
We have been and still are very open to discussion.
But the discussion that took place in the close Board mailing list led
us to think that a new effort to further develop sme was doomed to
failure.
We set some key points (close to upstream as much as possible, proven
technical base, well known and widely used components), we aimed at
some goals (simplicity, user and sysadmin friendliness, easy
troubleshooting) and tried to imagine where we wanted to go.
Now, we have reached a point were we have something working, and we
wanted to show our work.
It's an alpha release, it has bugs, it is subject to changes. But it's
modular, easy to customize and modify.
It's GPL, every issue is publicly discussed on redmine.
I wish everybody a happy new year.
--
Ciao,
Filippo
Hello Filippo,

Thank you for keeping an open mind and sharing your thoughts. I recall
the discussions in 2009 and share your frustration.

As I understand, you have a significant customer base in Italy and the
new version you and your team are working on is intended to service this
market. We are talking of a commercial venture, correct? All of this
development has to be paid for, how do you propose to recover costs?
Will there be any subscription fee for updates, or access to modules, or
any other cost-recovery mechanism? Would you care to share your
business model?

It would also be useful if you could share your vision in terms of the
existing user base of SME. How do you propose to address the following
topics:
- Mirrors? How many and where?
- Support?
- Forums?

Finally, how far are you prepared to compromise to accommodate the needs
of "low budget" users, in particular NGO, schools, non-profit
organisations etc? At the moment, SME is not very demanding. The
software can be installed on rather modest hardware and configured using
an old monitor thanks to a simple graphical interface and console. As a
result, there are quiet a few "older" servers out in the field still in
good working condition - replacing them to meet Nethserver H/W
requirements may not be feasible in the short term. Any plan for a 32
bits version?

Thanks.
chris
Filippo Carletti
2013-01-15 11:51:29 UTC
Permalink
We are talking of a commercial venture, correct? All of this development has
to be paid for, how do you propose to recover costs? Will there be any
subscription fee for updates, or access to modules, or any other
cost-recovery mechanism? Would you care to share your business model?
Nethesis is an established commercial venture. You could find more
information (in Italian) on our website, our blog or our pages in the
primary social networks.
We "sell" support in Italian language, by phone or email. We are
paying for development, money comes from support revenues.
We don't plan to charge for updates of packages. We plan to offer
subscriptions to services like remote backup, remote update,... (see
http://www.nethesis.it/en/monitoraggio-centralizzato-nethsecurity for
hints).
It would also be useful if you could share your vision in terms of the
existing user base of SME. How do you propose to address the following
- Mirrors? How many and where?
- Support?
- Forums?
We still have to think about those, but I don't think it will be hard
to setup some mirrors or forums. We are open to suggestions.
We quickly setup a mailing list using google groups.
We plan to spend the majority of working time developing the product
in the coming weeks.
After that, we will work on infrastructure.
Finally, how far are you prepared to compromise to accommodate the needs of
"low budget" users, in particular NGO, schools, non-profit organisations
etc? At the moment, SME is not very demanding. The software can be
installed on rather modest hardware and configured using an old monitor
thanks to a simple graphical interface and console. As a result, there are
quiet a few "older" servers out in the field still in good working condition
- replacing them to meet Nethserver H/W requirements may not be feasible in
the short term. Any plan for a 32 bits version?
NethServer is as demanding as Centos 6.3, not too far from SME. We
will support low budget users. NethServer console is simple,
server-manager requires a modern browser or a smartphone (btw, has
anybody tried using the server-manager with a smartphone?).
No plan for a 32 bit version, but it could be done (if not directly by
us, by a volunteer, following our documentation, see dev.nethesis.it).


--
Ciao,
Filippo
chris burnat
2013-01-15 22:24:48 UTC
Permalink
Thank you Filippo.
Additional questions if I may.

Nethesis is an established commercial venture. You could find more
information (in Italian) on our website, our blog or our pages in the
primary social networks.
We "sell" support in Italian language, by phone or email. We are
paying for development, money comes from support revenues.
We don't plan to charge for updates of packages. We plan to offer
subscriptions to services like remote backup, remote update,... (see
http://www.nethesis.it/en/monitoraggio-centralizzato-nethsecurity for
hints).

OK, so I move from SME to Nethesis. The ISO is basically CentOS minimal
+ nethserver-base + nethserver-httpd-admin (web UI). Any other
functionality requires modules, here is a tentative list of individual
module:

Users: nethserver-directory
DNS server and DHCP server: nethserver-dnsmasq and nethserver-hosts
HTTPD: nethserver-httpd (add also ibay httpd access)
Configuration can be done via template or /etc/httpd/conf.d
directory (preferred way)
Mail server: nethserver-mail-server (postifx + dovecot2)
Mail filtering: nethserver-mail-filter (amavis + spamassassin)
NTP server: netherver-ntp
Ibays: nethserver-ibays
SSH: ssh server is installed and accessibile by default,
nethserver-openssh add shell access for users
Samba: nethserver-samba
MySQL: nethserver-mysql (almost done)
Jabber: nethserver-ejabberd (almost done)
Printers: nethserver-cups (almost done)
Fax server: nethserver-hylafax (almost done)
Groupware: nethserver-sogo (work in progress). Porting Horde to
nethserver, should be a quick job.
Backup: working on. We plan to have two
Firewall module. Now we use CentOS standard but we plan to add a
more advanced firewall module like Shorewall.

a) Which of these are free and which of these need to be subscribe to?

b) These modules will be "bunched up" by category (Neth service, Neth
security etc), what will be the cost of annual subscription?

c) Any scope for a user to add customised modules, or for that matter
contributions as is the case with SME?

Thanks and regards
chris
Filippo Carletti
2013-01-16 14:24:46 UTC
Permalink
[Modules list]
Post by chris burnat
a) Which of these are free and which of these need to be subscribe to?
All free. See http://code.nethesis.it/.
Post by chris burnat
b) These modules will be "bunched up" by category (Neth service, Neth
security etc), what will be the cost of annual subscription?
You could create a custom cd that installs some packages and name it
as you like.
Subscription is not needed.
Post by chris burnat
c) Any scope for a user to add customised modules, or for that matter
contributions as is the case with SME?
Writing a contrib easily was a requirement.
We have some limited documentation:
https://dev.nethesis.it/projects/nethserver/wiki/Creating_web_UI_module


--
Ciao,
Filippo
chris burnat
2013-01-16 23:02:10 UTC
Permalink
This is good news.
Thanks Filippo.
chris
Post by Filippo Carletti
[Modules list]
Post by chris burnat
a) Which of these are free and which of these need to be subscribe to?
All free. See http://code.nethesis.it/.
Post by chris burnat
b) These modules will be "bunched up" by category (Neth service, Neth
security etc), what will be the cost of annual subscription?
You could create a custom cd that installs some packages and name it
as you like.
Subscription is not needed.
Post by chris burnat
c) Any scope for a user to add customised modules, or for that matter
contributions as is the case with SME?
Writing a contrib easily was a requirement.
https://dev.nethesis.it/projects/nethserver/wiki/Creating_web_UI_module
--
Ciao,
Filippo
Gordon Rowell
2013-01-14 20:42:29 UTC
Permalink
On Tue, Jan 15, 2013 at 3:49 AM, Filippo Carletti <
Post by Filippo Carletti
[English is not my native language, please bear it in mind if some of
my sentences may seem incorrect/inappropriate]
If my German and French were anywhere near your English, I would be
delighted.
Post by Filippo Carletti
I also don't understand why this effort wasn't put into SME
Server 8. Supporting both simply won't happen and the guaranteed
fragmentation of the community is sad.
Gordon, nice to hear from you.
I think that you (and many others reading here) are missing the
discussion that took place in September/October 2009 (more than three
years ago) in the "Board" mailing list.
There, we talked a lot about the future of SME and came to nothing.
Agreed. That exercise was a dismal failure. It should be completely
ignored. The only way SME Server will survive is for developer and tester
involvement.

The whole "board" concept was to provide a financial base for the
distribution. That failed, multiple times.
Post by Filippo Carletti
During the last years, I, Federico and Giacomo contributed to sme
providing bug fixing, patches, new features, testing and a bit of
support. Sometimes, we had our patches refused, especially for new
features. It forced us to fork some sme packages, we did that with
disappointment, we tried really hard not to fork, but we have a
business and our customers always have a point, so if the feature they
were asking for was useful we developed it, submitting a description
to contribs bugzilla, providing code and building packages if code was
approved.
We have been and still are very open to discussion.
Excellent.

Having patches refused and modified is a normal part of development. As
long as those discussions were openly tracked in Bugzilla, I think it's
perfectly reasonable to come to the decision that you need to fork
individual components for your use This, again, is normal. Ideally those
forks would be available in the contribs repository to simplify a merge at
some later stage if others felt your changes in direction were correct.

However, I am concerned that a wholesale rewrite of so many components has
been done and it will simply not be possible to reverse that without a huge
effort.

But the discussion that took place in the close Board mailing list led
Post by Filippo Carletti
us to think that a new effort to further develop sme was doomed to
failure.
That is quite sad, but I understand this view. I think you should
completely ignore the failure of the board (and any attempts to do such
again) and concentrate on the technical aspects.
Post by Filippo Carletti
We set some key points (close to upstream as much as possible, proven
technical base, well known and widely used components), we aimed at
some goals (simplicity, user and sysadmin friendliness, easy
troubleshooting) and tried to imagine where we wanted to go.
These sounds like good goals, and, I believe are totally in line with SME
Server development over the years. We always avoided forks from upstream
unless there was no option (templates helped immensely here) and made some
choices based on best of breed at the time which may be different today
(e.g qmail saved us from at least one remote sendmail exploits. I'd choose
Postfix today, but I actually think that change is neutral from a server
capability p.o.v.).
Post by Filippo Carletti
Now, we have reached a point were we have something working, and we
wanted to show our work.
It's an alpha release, it has bugs, it is subject to changes. But it's
modular, easy to customize and modify.
It's GPL, every issue is publicly discussed on redmine.
That's excellent. I wish you all the best.

Gordon
Post by Filippo Carletti
I wish everybody a happy new year.
--
Ciao,
Filippo
Hsing-Foo Wang
2013-01-14 22:54:04 UTC
Permalink
Its 'accept everything
and bounce bad addresses later' policy forced me to drop SME as mail
servers long ago. Maybe that was fixed eventually...
So we are able to harvest some possible development targets during the last
few days.

@Gordon, Agree with the focus on patching, testing, testing and testing
@Charlie, I hope perl devs will speak out loud and pick up the challenge

Filippo, would you care to join efforts and see how you can share your
team, and not lose your business focus/objectives?

If yes, we will work out a way to get that established to the benefit of
'us' all.

@all, could you please provide your 1 liner 'support or not' to this
thread? Your short constructive feedback would provide an idea about the
current willingness to add to our efforts.

Maybe like:

---
Yes I support SME Server.

Name: John Doe
Skills: Perl, PHP, Bash, qmail, etc.etc.etc
Time commitment: 5 hours per week
Role commitments: leader/worker/assistant
---

We have been here before, but this time, no nonsense...
Filippo Carletti
2013-01-15 12:08:28 UTC
Permalink
Post by Gordon Rowell
The whole "board" concept was to provide a financial base for the
distribution. That failed, multiple times.
I think that we are offering a financial base, in a sense. :-)
Post by Gordon Rowell
Having patches refused and modified is a normal part of development. As long
as those discussions were openly tracked in Bugzilla, I think it's perfectly
reasonable to come to the decision that you need to fork individual
components for your use This, again, is normal. Ideally those forks would be
available in the contribs repository to simplify a merge at some later stage
if others felt your changes in direction were correct.
The forked packages are available in our repositories and, in patch
form, in bugzilla.
Post by Gordon Rowell
However, I am concerned that a wholesale rewrite of so many components has
been done and it will simply not be possible to reverse that without a huge
effort.
I don't want to sound bold, but if everybody agrees that changes are
needed we will obviously make our part of the job.
We want a good product.
Post by Gordon Rowell
Post by Filippo Carletti
But the discussion that took place in the close Board mailing list led
us to think that a new effort to further develop sme was doomed to
failure.
That is quite sad, but I understand this view. I think you should completely
ignore the failure of the board (and any attempts to do such again) and
concentrate on the technical aspects.
That's what we did: we developed NethServer. We planned to share it
with this community much earlier, but I was sidetracked by other
problems while development continued.
I'm sorry for my mistakes, I'm listening to every suggestion or
complaint, trying to fix the errors.

I really appreciate your comments, thank you Gordon.

--
Ciao,
Filippo
Steven Santos
2013-01-15 13:06:35 UTC
Permalink
Flippo,

First, I want to say thank you. Your work and willingness to work
with the community is huge.

I look forward to more developments with this, as I think it might be
the future of SME
---
Steven Santos
Director
Simply Circus, Inc.
86 Los Angeles Street
Newton, MA 02458

P: 617-527-0667
F: 617-934-1870
E: ***@SimplyCircus.com


On Tue, Jan 15, 2013 at 7:08 AM, Filippo Carletti
Post by Filippo Carletti
Post by Gordon Rowell
The whole "board" concept was to provide a financial base for the
distribution. That failed, multiple times.
I think that we are offering a financial base, in a sense. :-)
Post by Gordon Rowell
Having patches refused and modified is a normal part of development. As long
as those discussions were openly tracked in Bugzilla, I think it's perfectly
reasonable to come to the decision that you need to fork individual
components for your use This, again, is normal. Ideally those forks would be
available in the contribs repository to simplify a merge at some later stage
if others felt your changes in direction were correct.
The forked packages are available in our repositories and, in patch
form, in bugzilla.
Post by Gordon Rowell
However, I am concerned that a wholesale rewrite of so many components has
been done and it will simply not be possible to reverse that without a huge
effort.
I don't want to sound bold, but if everybody agrees that changes are
needed we will obviously make our part of the job.
We want a good product.
Post by Gordon Rowell
Post by Filippo Carletti
But the discussion that took place in the close Board mailing list led
us to think that a new effort to further develop sme was doomed to
failure.
That is quite sad, but I understand this view. I think you should completely
ignore the failure of the board (and any attempts to do such again) and
concentrate on the technical aspects.
That's what we did: we developed NethServer. We planned to share it
with this community much earlier, but I was sidetracked by other
problems while development continued.
I'm sorry for my mistakes, I'm listening to every suggestion or
complaint, trying to fix the errors.
I really appreciate your comments, thank you Gordon.
--
Ciao,
Filippo
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Charlie Brady
2013-01-15 20:21:34 UTC
Permalink
Post by Filippo Carletti
Post by Gordon Rowell
The whole "board" concept was to provide a financial base for the
distribution. That failed, multiple times.
I think that we are offering a financial base, in a sense. :-)
A financial base, but to Nethserver, not to SME server.

Nethserver is a new and different product, which just shares some design
concepts and some code with SME server. IIUC, there is no migration path
from one to the other. Nethserver was not built by making changes to SME
server, but by starting afresh and just cherry picking some code. [I hope
I am not misinterpreting, and apologies if I have.]
Post by Filippo Carletti
Post by Gordon Rowell
Having patches refused and modified is a normal part of development. As long
as those discussions were openly tracked in Bugzilla, I think it's perfectly
reasonable to come to the decision that you need to fork individual
components for your use This, again, is normal. Ideally those forks would be
available in the contribs repository to simplify a merge at some later stage
if others felt your changes in direction were correct.
The forked packages are available in our repositories and, in patch
form, in bugzilla.
I'm not sure that you and Gordon were talking about the same set of forks.
I doubt that all of your code which is different to SME server code is in
patch form, in bugzilla.
Post by Filippo Carletti
Post by Gordon Rowell
However, I am concerned that a wholesale rewrite of so many components has
been done and it will simply not be possible to reverse that without a huge
effort.
I don't want to sound bold, but if everybody agrees that changes are
needed we will obviously make our part of the job.
I doubt that everyone will agree about very much :-)

It also remains true, I think, that reversing the extensive changes you
have made would be a huge effort.
Post by Filippo Carletti
We want a good product.
That's great. Whether that product is just great for your business, or it
can fully supplant SME server without excessive disruption or loss of
utility remains to be seen.

I don't like to be boring, but I will again raise PHP as a critical issue.
I will no longer run PHP applications on my server. I will certainly not
run PHP applications which either directly or indirectly have privileged
administrative access.

I know you have already put a lot of work into your UI code - 32000 lines
in just the base install packages. But I think that decision was a serious
mistake. But Nethserver is your baby, and it is your business. I hope you
have good fortune.
Post by Filippo Carletti
Post by Gordon Rowell
Post by Filippo Carletti
But the discussion that took place in the close Board mailing list led
us to think that a new effort to further develop sme was doomed to
failure.
That is quite sad, but I understand this view. I think you should completely
ignore the failure of the board (and any attempts to do such again) and
concentrate on the technical aspects.
That's what we did: we developed NethServer.
Gordon was only suggesting that you ignore the discussions about an SME
server board. He wasn't suggesting you should ignore the SME server
developer community.
Post by Filippo Carletti
We planned to share it
with this community much earlier, but I was sidetracked by other
problems while development continued.
I'm sorry for my mistakes, I'm listening to every suggestion or
complaint, trying to fix the errors.
I really appreciate your comments, thank you Gordon.
--
Ciao,
Filippo
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Greg Zartman
2013-01-14 18:11:15 UTC
Permalink
Post by Gordon Rowell
As a long time lurker since I stopped being able to commit time as an
active developer here, I find this announcement both exciting and sad. It
is both a new beginning and a death, as a fork of SME Server (which this
is), means that one or both will die. While I applaud the move to a new
base, I'm not sure of the rationale behind replacing components that have
worked well for many years. I also don't understand why this effort wasn't
put into SME Server 8. Supporting both simply won't happen and the
guaranteed fragmentation of the community is sad.
I've been somewhat of a lurker these past 2-3 years as well, due to my
change in job responsibilities.

I agree with Gordon here. This is both exciting and sad. It is exciting
to see someone picking up SME and making a real effort to move it forward.
However, it is sad to see this happen without some collaboration with the
developers at contribs.org who have put so much into the current
architecture nor the community at large. There is a huge knowledge base
here that is being thrown out the window with this effort.

It is exciting to see the Neth folks looking at integrating new
technologies into SME. Some of the current technology hasn't been changed
much in over a decade (e.g. qmail, samba). This is something I've wanted
to see for some time and I tried as much as I could, and time would allow,
to do things where my experience could help (Samba). Personally, I'm
slowly moving away from SME because some of these technologies are getting
long in the tooth and lack the functionality of more modern code. For
example, I've moved my company email server to google apps because SME is
soooo far behind in terms of content filtering, mail sorting, etc.

I hope the Neth folks take note of the concern being voiced by the
community about collaboration, otherwise I agree with Gordon. One of both
of these distros is going to die. I for one would not deploy Nethserver
in my shop. The Neth folks are doing some things that raise real concerns
for me, which Charlie and others have also voice. I'd rather go with a
home brew Centos solution or similar, to maintain security and stability.

Greg Zartman
Michael Doerner - TechnologyWise
2013-01-13 10:08:54 UTC
Permalink
Post by Stefano Zamboni
Post by Graeme Robinson
agree, with an exclamation mark. Postfix over qmail? I don't think so.
I think we all are missing the point.. this is not a qmail vs postfix
question..
AFAIK SME development is almost stopped
CentOS5 is getting old, 6 is already out, stable..
do WE have resources to move SME to CentOS6?
should we waste time in "$our_favourite_daemon" vs "$new_daemon" war?
we have knowledge, know how.. since we are not asked to change
distribution we should be able to move to the new, aren't we?
we have an opportunity here: we have a company that support
development.. can we, really, miss this train?
I don't think so
I fully agree with Stefano.

The " qmail vs postfix question.." is minor, down at the bottom end of
the list. The questions of survival for SME have never been solved and
here seems to be a chance to move on.

We can't thank enough the remaining developers, contributors and other
helpers to come to this point for SMEserver but the chance for a future
of the brilliant ideas of events, templating system and configuration
database seems very unlikely with SME.

Michael
stephane de Labrusse
2013-01-13 10:25:15 UTC
Permalink
I make a donation paypal to the sme server project for several years,
as I'm not a developer this is my personal way to participate in the
project.

Do you want to implement a system of monthly contribution, I can not
put a fortune, but small streams become rivers.

I will continue to give money to contribs.org, this even if nobody
explains how money is spent ..... This is a pity.

- --

Cordialement

Stéphane de Labrusse
Président de l'aru2L http://www.aru2L.org
Membre de l'ApriL http://www.apriL.org


PORTABLE : 06 29 19 12 99
DOMICILE : 05 65 78 90 72
WEB : http://geekeries.de-labrusse.fr
Post by Michael Doerner - TechnologyWise
Post by Stefano Zamboni
Post by Graeme Robinson
agree, with an exclamation mark. Postfix over qmail? I don't think so.
I think we all are missing the point.. this is not a qmail vs
postfix question..
AFAIK SME development is almost stopped
CentOS5 is getting old, 6 is already out, stable..
do WE have resources to move SME to CentOS6?
should we waste time in "$our_favourite_daemon" vs "$new_daemon" war?
we have knowledge, know how.. since we are not asked to change
distribution we should be able to move to the new, aren't we?
we have an opportunity here: we have a company that support
development.. can we, really, miss this train?
I don't think so
I fully agree with Stefano.
The " qmail vs postfix question.." is minor, down at the bottom end
of the list. The questions of survival for SME have never been
solved and here seems to be a chance to move on.
We can't thank enough the remaining developers, contributors and
other helpers to come to this point for SMEserver but the chance
for a future of the brilliant ideas of events, templating system
and configuration database seems very unlikely with SME.
Michael
_______________________________________________ Server Development
Discussion To unsubscribe, e-mail
http://lists.contribs.org/mailman/public/devinfo/
Stefano Zamboni
2013-01-13 10:44:50 UTC
Permalink
@filippo:

- do we really need graphical setup? can't we move back to text-only setup?
- why google dns?
- if I delete google dns ip's, I'm not able to update the server.. since
I'm in server-only mode and I set my gateway correctly, why nethserver
is unable to resolve names?
- re-applying dns ips, I'm unable to update the server from your repo..
it seems like something is wrong, all your packages return 404
- packages management page is nice but If I choose to see 25 or more
elements per page, I can't see the link to move to another page (firefox
18 on 15.6" wide laptop screen)
- after updating the server, and after updating the kernel packages, I'm
not requested to reboot..
- after trying to update the server and no packages from nethserver repo
found (see above), how I can install/update them? yum update returns "no
packages found" a "youm clena all" doesn't solve the issue
- why move server-manager files to /usr/share?

IMHO moving to new server-manager programming language and/or new
daemons should not means change in core structure.. if you (nethesis)
wish to have help form the core SME developers, you should not change
everything

Ciao

Stefano
chris burnat
2013-01-13 11:14:00 UTC
Permalink
Post by Stefano Zamboni
- do we really need graphical setup? can't we move back to text-only setup?
- why google dns?
- if I delete google dns ip's, I'm not able to update the server..
since I'm in server-only mode and I set my gateway correctly, why
nethserver is unable to resolve names?
- re-applying dns ips, I'm unable to update the server from your
repo.. it seems like something is wrong, all your packages return 404
- packages management page is nice but If I choose to see 25 or more
elements per page, I can't see the link to move to another page
(firefox 18 on 15.6" wide laptop screen)
- after updating the server, and after updating the kernel packages,
I'm not requested to reboot..
- after trying to update the server and no packages from nethserver
repo found (see above), how I can install/update them? yum update
returns "no packages found" a "youm clena all" doesn't solve the issue
- why move server-manager files to /usr/share?
IMHO moving to new server-manager programming language and/or new
daemons should not means change in core structure.. if you (nethesis)
wish to have help form the core SME developers, you should not change
everything
Ciao
Stefano
Short comments:
- A proper bugtracker is mandatory, please consider Bugzilla - there are
possibly 10 bugs report in this post which could have been opened.
- I suspect that you will experience some resistance using Google Group
with a few of us, suggest you implement a mailing list system similar to
what we use at Contribs (i.e. the format used in this instance).
- I hope the SME admin console is yet to come, makes it difficult doing
things at the moment, or am I missing something?
- Forcing "it" keyboard is a real pain, especially since you do not
provide the usual installation options we have with SME during installation.
- When doing text only setup (preferred method), CD test fails, md5 does
not appear to be included in ISO.

Casual observation: If it ain't broke, don't fix/change it [old
proverb]. I suspect the new team will have enough to do without
re-inventing the wheel.
Charlie Brady
2013-01-13 19:08:31 UTC
Permalink
Post by chris burnat
- A proper bugtracker is mandatory, please consider Bugzilla - there are
possibly 10 bugs report in this post which could have been opened.
There is a bug tracker, and I'm sure that Filippo is familiar with
Bugzilla.

https://dev.nethesis.it/projects/nethserver/issues

I'm not very intimate with the redmine project management system, but I
suspect that Bugzilla might be a backward step.

http://www.redmine.org/

I hope that Filippo has patched his Ruby-on-Rails installation:

https://community.rapid7.com/community/metasploit/blog/2013/01/09/serialization-mischief-in-ruby-land-cve-2013-0156
chris burnat
2013-01-14 05:36:16 UTC
Permalink
Ooops....
Thanks Charlie.
chris
Post by Charlie Brady
Post by chris burnat
- A proper bugtracker is mandatory, please consider Bugzilla - there are
possibly 10 bugs report in this post which could have been opened.
There is a bug tracker, and I'm sure that Filippo is familiar with
Bugzilla.
https://dev.nethesis.it/projects/nethserver/issues
I'm not very intimate with the redmine project management system, but I
suspect that Bugzilla might be a backward step.
http://www.redmine.org/
https://community.rapid7.com/community/metasploit/blog/2013/01/09/serialization-mischief-in-ruby-land-cve-2013-0156
Giacomo Sanchietti
2013-01-14 16:00:23 UTC
Permalink
Hi,
my name is Giacomo Sanchietti.
I'm working in Nethesis as part of the SME Server development group. I'm
also one of the main developer of NethServer.
Post by chris burnat
- A proper bugtracker is mandatory, please consider Bugzilla - there are
possibly 10 bugs report in this post which could have been opened.
The bug tracker is now open and is available at:
https://dev.nethesis.it/projects/nethserver/issues
There is a tracker for each subproject.

You can register here:
https://dev.nethesis.it/account/register
Post by chris burnat
- I suspect that you will experience some resistance using Google Group
with a few of us, suggest you implement a mailing list system similar to
what we use at Contribs (i.e. the format used in this instance).
You don't need a Google account to use Google Group. We've added more
info here: http://nethserver.nethesis.it/index.php?id=support

By the way, if using a Google Group is a problem for many users, we can
setup a ML.
Post by chris burnat
- I hope the SME admin console is yet to come, makes it difficult doing
things at the moment, or am I missing something?
There is a minimal 'console' which is used on the first boot, and can be
launched with the 'console' command. But we plan to replace the text
console with a web-based one. Maybe the best choice is to have both.
Post by chris burnat
- Forcing "it" keyboard is a real pain, especially since you do not
provide the usual installation options we have with SME during
installation.
You're right. I will add a keyboard option to the alpha2 iso:
https://dev.nethesis.it/issues/1620

In the meantime you can use the traditional CentOS installation and
install NethServer via yum. Otherwise you can modify the kickstart file
and place it on a local http server.
Post by chris burnat
- When doing text only setup (preferred method), CD test fails, md5 does
not appear to be included in ISO.
Automatic installation do not perform any check on the iso. The test
fails because we added some packages and kickstart file to the original
CentOS minimal iso (see
http://code.nethesis.it/?p=nethserver-cdrom.git;a=tree;h=refs/heads/master;hb=master).
If you wish to check the image, please do it manually downloading the
md5 and using the md5sum command.
--
Giacomo Sanchietti
Charlie Brady
2013-01-14 16:27:33 UTC
Permalink
Post by chris burnat
- When doing text only setup (preferred method), CD test fails, md5 does
not appear to be included in ISO.
Automatic installation do not perform any check on the iso. The test fails
because we added some packages and kickstart file to the original CentOS
minimal iso (see
http://code.nethesis.it/?p=nethserver-cdrom.git;a=tree;h=refs/heads/master;hb=master).
If you wish to check the image, please do it manually downloading the md5 and
using the md5sum command.
Use 'implantisomd5' to update the md5 sum in the iso file.
Giacomo Sanchietti
2013-01-14 16:58:04 UTC
Permalink
Post by Charlie Brady
Use 'implantisomd5' to update the md5 sum in the iso file.
You're right: it works like a charm.

Added a bug for the next alpha release.

Thank you
--
Giacomo Sanchietti
chris burnat
2013-01-14 20:40:06 UTC
Permalink
Hello Giacomo,

Good to hear from you, thanks for clarifications.
I believe that adding a full console (graphical/text) with similar
functionality as the current SME version would be a good idea.
Regards
chris

------------------------------------------------------------
Post by Giacomo Sanchietti
Hi,
my name is Giacomo Sanchietti.
I'm working in Nethesis as part of the SME Server development group.
I'm also one of the main developer of NethServer.
Post by chris burnat
- A proper bugtracker is mandatory, please consider Bugzilla - there are
possibly 10 bugs report in this post which could have been opened.
https://dev.nethesis.it/projects/nethserver/issues
There is a tracker for each subproject.
https://dev.nethesis.it/account/register
Post by chris burnat
- I suspect that you will experience some resistance using Google Group
with a few of us, suggest you implement a mailing list system similar to
what we use at Contribs (i.e. the format used in this instance).
You don't need a Google account to use Google Group. We've added more
info here: http://nethserver.nethesis.it/index.php?id=support
By the way, if using a Google Group is a problem for many users, we
can setup a ML.
Post by chris burnat
- I hope the SME admin console is yet to come, makes it difficult doing
things at the moment, or am I missing something?
There is a minimal 'console' which is used on the first boot, and can
be launched with the 'console' command. But we plan to replace the
text console with a web-based one. Maybe the best choice is to have both.
Post by chris burnat
- Forcing "it" keyboard is a real pain, especially since you do not
provide the usual installation options we have with SME during installation.
https://dev.nethesis.it/issues/1620
In the meantime you can use the traditional CentOS installation and
install NethServer via yum. Otherwise you can modify the kickstart
file and place it on a local http server.
Post by chris burnat
- When doing text only setup (preferred method), CD test fails, md5 does
not appear to be included in ISO.
Automatic installation do not perform any check on the iso. The test
fails because we added some packages and kickstart file to the
original CentOS minimal iso (see
http://code.nethesis.it/?p=nethserver-cdrom.git;a=tree;h=refs/heads/master;hb=master).
If you wish to check the image, please do it manually downloading the
md5 and using the md5sum command.
Giacomo Sanchietti
2013-01-14 16:31:51 UTC
Permalink
Hi,
Post by Stefano Zamboni
- do we really need graphical setup? can't we move back to text-only setup?
Do you mean the anaconda installation? In the kickstart file there is
the 'text' option which should force the text installation.
But anaconda seems to ignore the parameter if the system has more than
512MB of RAM. I will search for further documentation.
Post by Stefano Zamboni
- why google dns?
Why not? :) We can choose any DNS as default, like OpenDNS.
Post by Stefano Zamboni
- if I delete google dns ip's, I'm not able to update the server.. since
I'm in server-only mode and I set my gateway correctly, why nethserver
is unable to resolve names?
Because, like CentOS, NethServer is not a DNS server (and this is why we
have default external DNS). If you need to use NethServer as a DNS
server, you have to install nethserver-dnsmasq module.
Post by Stefano Zamboni
- re-applying dns ips, I'm unable to update the server from your repo..
it seems like something is wrong, all your packages return 404
I can't reproduce the problem. Can you try again? Then tell me date and
time and I will check the logs.
Post by Stefano Zamboni
- packages management page is nice but If I choose to see 25 or more
elements per page, I can't see the link to move to another page (firefox
18 on 15.6" wide laptop screen)
We use a standard jquery table plugin. You should see 'next' label on
the bottom right corner of the table. Also you can use the search field
(on the top right corner)
Post by Stefano Zamboni
- after updating the server, and after updating the kernel packages, I'm
not requested to reboot..
That's right, we are trying to minimize the needed reboots.
Post by Stefano Zamboni
- after trying to update the server and no packages from nethserver repo
found (see above), how I can install/update them? yum update returns "no
packages found" a "youm clena all" doesn't solve the issue
Here the direct link to the repo:
http://pulp.nethesis.it/nethserver/6.3/base/x86_64/
Please try again and let me know.
Post by Stefano Zamboni
- why move server-manager files to /usr/share?
Because it's not written in perl anymore, so we just moved to another
directory. We followed the FHS, but are there any other requirements
about the directory structure?
Post by Stefano Zamboni
IMHO moving to new server-manager programming language and/or new
daemons should not means change in core structure.. if you (nethesis)
wish to have help form the core SME developers, you should not change
everything
We didn't change everything, but we spent a lot of time to clean-up
packages removing old code and trying to keep various modules loose coupled.

Feel free to ask question about all changes we've made and I will glad
to answer why we choose that path.

This is an alpha release, we are open to changes.
--
Giacomo Sanchietti
Stefano Zamboni
2013-01-14 18:32:53 UTC
Permalink
Post by Giacomo Sanchietti
Hi,
Hi Giacomo :-)
Post by Giacomo Sanchietti
Post by Stefano Zamboni
- do we really need graphical setup? can't we move back to text-only setup?
Do you mean the anaconda installation? In the kickstart file there is
the 'text' option which should force the text installation.
But anaconda seems to ignore the parameter if the system has more than
512MB of RAM. I will search for further documentation.
yes, I was referring to anaconda
Post by Giacomo Sanchietti
Post by Stefano Zamboni
- why google dns?
Why not? :) We can choose any DNS as default, like OpenDNS.
Post by Stefano Zamboni
- if I delete google dns ip's, I'm not able to update the server.. since
I'm in server-only mode and I set my gateway correctly, why nethserver
is unable to resolve names?
Because, like CentOS, NethServer is not a DNS server (and this is why
we have default external DNS). If you need to use NethServer as a DNS
server, you have to install nethserver-dnsmasq module.
IMHO (reading other mails too) you are following the same approach of
zentyal.. a distro with some pagkages.. i.e. install centos then make it
become SME..
sincerely.. I don't like it
since I'm installing a SERVER distro, I expect I can find all the
minimal tools when I finish setup
so, I expect to find samba, dns server.. I, as an experienced SME user,
expect to find everything I find in SME.. it's up to me to disable what
I don't need
Post by Giacomo Sanchietti
Post by Stefano Zamboni
- re-applying dns ips, I'm unable to update the server from your repo..
it seems like something is wrong, all your packages return 404
I can't reproduce the problem. Can you try again? Then tell me date
and time and I will check the logs.
I will reinstall this evening, try again and let you know
Post by Giacomo Sanchietti
We use a standard jquery table plugin. You should see 'next' label on
the bottom right corner of the table. Also you can use the search
field (on the top right corner)
that's the issue.. I can't see next label..
and search is useful if I know what to search..
Post by Giacomo Sanchietti
Post by Stefano Zamboni
- after updating the server, and after updating the kernel packages, I'm
not requested to reboot..
That's right, we are trying to minimize the needed reboots.
let me understand: updating kernel does not require reboot? if so, how
did you achieve it?
Post by Giacomo Sanchietti
Post by Stefano Zamboni
- after trying to update the server and no packages from nethserver repo
found (see above), how I can install/update them? yum update returns "no
packages found" a "youm clena all" doesn't solve the issue
http://pulp.nethesis.it/nethserver/6.3/base/x86_64/
Please try again and let me know.
I will, see above
Post by Giacomo Sanchietti
We didn't change everything, but we spent a lot of time to clean-up
packages removing old code and trying to keep various modules loose coupled.
Feel free to ask question about all changes we've made and I will glad
to answer why we choose that path.
This is an alpha release, we are open to changes.
ok.. can we change the point of view? could you make a list of the main
changes from SME standards?
in this way we don't waste time searching things :-)

TIA

Ciao
Stefano
Giacomo Sanchietti
2013-01-15 13:24:47 UTC
Permalink
Post by Stefano Zamboni
IMHO (reading other mails too) you are following the same approach of
zentyal.. a distro with some pagkages.. i.e. install centos then make it
become SME..
sincerely.. I don't like it
since I'm installing a SERVER distro, I expect I can find all the
minimal tools when I finish setup
so, I expect to find samba, dns server.. I, as an experienced SME user,
expect to find everything I find in SME.. it's up to me to disable what
I don't need
We use a "keep it simple" approach to minimize resources and hard disk
space occupation.
But you can create a personalized kickstart file which can install all
desired stuff automatically. You could also have "server profiles", like
file server, friewall, virtualization box etc.
Post by Stefano Zamboni
that's the issue.. I can't see next label..
and search is useful if I know what to search..
I just tested with Firefox 18 and it works well.
Can you try with another browser? Or you can disable javascript.

The package manager shows a list of yum group. Just as a workaround, you
can open NethServer.repo file and add 'enablegroups=0' to centos-base
repository, then clean the yum cache with 'yum clean all'.
Post by Stefano Zamboni
Post by Giacomo Sanchietti
That's right, we are trying to minimize the needed reboots.
let me understand: updating kernel does not require reboot? if so, how
did you achieve it?
The reboot is necessary only if you want to load the new kernel. Often
this is not needed in production environments.
But this is only an alpha release, in future versions we will add a
notification when the reboot is recommended.
Post by Stefano Zamboni
ok.. can we change the point of view? could you make a list of the main
changes from SME standards?
in this way we don't waste time searching things :-)
Sure, you're right.

We created a new page with all technical differences between NethServer
and SME:
http://nethserver.nethesis.it/index.php?id=development_nethserver-differences
--
Giacomo Sanchietti
Charlie Brady
2013-01-15 21:12:29 UTC
Permalink
We created a new page with all technical differences between NethServer and
http://nethserver.nethesis.it/index.php?id=development_nethserver-differences
Thanks.

There's at least one technical difference I see missing there which I
consider significant. Nethserver does not install a single drive as a
degraded RAID1 mirror. So it is not possible to upgrade a server from
single drive to RAID1 pair without re-install.

I haven't looked through your changelogs or your bug tracker, but I wonder
whether you have written down the reasons for the various technical
decisions you have made, e.g. amavis v qpsmtpd, google dns vs local
dnscache, kickstart vs modified anaconda, the rationale for the event
queue, etc.
Giacomo Sanchietti
2013-01-16 16:39:04 UTC
Permalink
Post by Charlie Brady
There's at least one technical difference I see missing there which I
consider significant. Nethserver does not install a single drive as a
degraded RAID1 mirror. So it is not possible to upgrade a server from
single drive to RAID1 pair without re-install.
Yes, it's true. We can change the default or add an option like
'raid=force' to force the raid creation. It's quite simple to modify the
kickstart file.
Post by Charlie Brady
I haven't looked through your changelogs or your bug tracker, but I wonder
whether you have written down the reasons for the various technical
decisions you have made,
Some discussion haven't been documented, but in most cases the choice
was always to follow upstream decisions and use well tested solutions
(this is the case of postifix vs amavis).
Post by Charlie Brady
google dns vs local dnscache
We want a modular system: not all users need an authoritative dns.
And dnscache, is simple and widely used.
Post by Charlie Brady
kickstart vs modified anaconda,
This approach has many advantages:
* no need to patch anaconda and add possible bugs
* very easy to create a new iso when a new CentOS version will be released
* Kickstart is a standard and powerful tool. With kickstart anyone can
create a custom installation, publishing a new kickstart file for
example on an http server.
* We can create server 'profiles': firewall, kvm box, mail server, etc.
Post by Charlie Brady
the rationale for the event queue, etc.
The basic idea is to push into a queue any event signaled by RPM
scriptlets and execute them all only at the end of the yum transaction.
Using the queue we can assure all actions in the %post sections are
executed in the right order respecting dependencies between NethServer
modules.

See also:
https://dev.nethesis.it/projects/nethserver/wiki/NethServer_yum_plugin

If needed, I'll be glad to give you more information.
--
Giacomo Sanchietti
Shad L. Lords
2013-01-16 16:55:28 UTC
Permalink
On Wed, Jan 16, 2013 at 9:39 AM, Giacomo Sanchietti <
Post by Charlie Brady
There's at least one technical difference I see missing there which I
Post by Charlie Brady
consider significant. Nethserver does not install a single drive as a
degraded RAID1 mirror. So it is not possible to upgrade a server from
single drive to RAID1 pair without re-install.
Yes, it's true. We can change the default or add an option like
'raid=force' to force the raid creation. It's quite simple to modify the
kickstart file.
You can't do this option in the kickstart file. You need to modify
anaconda in order to make it understand degraded arrays. Looking at the
code you can't even upgrade a system unless the array is complete. The way
the RH6 does its raid is completely different to the way RH5 anaconda
worked. This isn't a simple thing of looking at how we did things for SME8
and apply them to RH6 anaconda.

the rationale for the event queue, etc.
Post by Charlie Brady
The basic idea is to push into a queue any event signaled by RPM
scriptlets and execute them all only at the end of the yum transaction.
Using the queue we can assure all actions in the %post sections are
executed in the right order respecting dependencies between NethServer
modules.
If you look at the sme yum plugin you can easily see what rpms are being
installed/removed. You could have just used this list instead of requiring
every post script to add and event to the queue. It could have been as
simple as looking for an update directory or event to exist for the
installed rpm.

-Shad
Giacomo Sanchietti
2013-01-17 10:08:06 UTC
Permalink
Post by Shad L. Lords
You can't do this option in the kickstart file. You need to modify
anaconda in order to make it understand degraded arrays. Looking at the
code you can't even upgrade a system unless the array is complete. The
way the RH6 does its raid is completely different to the way RH5
anaconda worked. This isn't a simple thing of looking at how we did
things for SME8 and apply them to RH6 anaconda.
I played a bit with kickstart and I can fool anaconda forcing to install
on a degraded array. But after reboot grub is not correctly installed.
Actually I don't have any spare time to dive into the problem, but
probably we can find a workaround (maybe using the %post section script).

Here's the code.

In %pre script:

mnt/runtime/usr/sbin/sfdisk /dev/$target1 -uM << EOF
,500
,
;
EOF

/mnt/runtime/usr/sbin/kpartx /dev/${target1}
echo y | /mnt/runtime/usr/sbin/mdadm --create /dev/md1 --level=1 -f -n1
/dev/${target1}1
echo y | /mnt/runtime/usr/sbin/mdadm --create /dev/md2 --level=1 -f -n1
/dev/${target1}2


And in the partition configuration section:

raid /boot --device=md1 --fstype='ext3' --level=1 --useexisting
raid pv.1 --device=md2 --fstype='physical volume (LVM)' --level=1
--useexisting
--
Giacomo Sanchietti
Shad L. Lords
2013-01-17 16:45:28 UTC
Permalink
On Thu, Jan 17, 2013 at 3:08 AM, Giacomo Sanchietti <
Post by Giacomo Sanchietti
I played a bit with kickstart and I can fool anaconda forcing to install
on a degraded array. But after reboot grub is not correctly installed.
Actually I don't have any spare time to dive into the problem, but probably
we can find a workaround (maybe using the %post section script).
Here's the code.
mnt/runtime/usr/sbin/sfdisk /dev/$target1 -uM << EOF
,500
,
;
EOF
/mnt/runtime/usr/sbin/kpartx /dev/${target1}
echo y | /mnt/runtime/usr/sbin/mdadm --create /dev/md1 --level=1 -f -n1
/dev/${target1}1
echo y | /mnt/runtime/usr/sbin/mdadm --create /dev/md2 --level=1 -f -n1
/dev/${target1}2
Where is target1 being set? Also, you can't just append a 1 or 2 to a
device to get the 1st and 2nd partition. This does work in many cases but
fails for at least the cciss devices.

-Shad
Stefano Zamboni
2013-01-17 16:49:55 UTC
Permalink
Post by Shad L. Lords
On Thu, Jan 17, 2013 at 3:08 AM, Giacomo Sanchietti
I played a bit with kickstart and I can fool anaconda forcing to
install on a degraded array. But after reboot grub is not
correctly installed. Actually I don't have any spare time to dive
into the problem, but probably we can find a workaround (maybe
using the %post section script).
I'm trying to reinstall Nethserver on the same VM I used some days ago..
setup routine did not see previous setup and formatted the hd..

I know we are in alpha stage but.... at least a "do you want to
proceed?" :-)

S.
Filippo Carletti
2013-01-17 16:53:05 UTC
Permalink
Post by Shad L. Lords
This does work in many cases but
fails for at least the cciss devices.
I tried a centos6 install on cciss many months ago, but I recall that
I was surprised to see that devices where sd*.
Are you sure that the old c0d0p0 format is still current on 6.x?

--
Ciao,
Filippo
Pascal Schirrmann
2013-01-17 19:52:27 UTC
Permalink
Post by Filippo Carletti
Post by Shad L. Lords
This does work in many cases but
fails for at least the cciss devices.
I tried a centos6 install on cciss many months ago, but I recall that
I was surprised to see that devices where sd*.
Are you sure that the old c0d0p0 format is still current on 6.x?
Hi,

Actually not anymore. The Compad/HP raid Array are now supported with
the standard scsi process in the last kernels.
Aside note : HP don't recommend too much to use the kernel driver... At
work, we use the upstream source (RedHat) and until now, we are using
the standard kernel driver, but we have a very limited set of physical
RedHat 6 systems. None of then beingused extensively on the local drives.

Pascal
Shad L. Lords
2013-01-17 19:56:44 UTC
Permalink
It isn't just the cciss. Any device that ends in a number will follow the
device#p# instead of just appending the # to the end of the device name. I
know the cciss used to behave this way and I've seen it happen on other
devices before on RH6. I just can't recall off the top of my head what
they were. If I find more information about them I'll pass it along.

-Shad
Giacomo Sanchietti
2013-01-17 16:55:14 UTC
Permalink
Post by Shad L. Lords
Where is target1 being set? Also, you can't just append a 1 or 2 to a
device to get the 1st and 2nd partition. This does work in many cases
but fails for at least the cciss devices.
target1 variable is set above (not reported here), it's just the disk
device name.

The code is only a proof of concept.

The current kickstart file is here :
http://code.nethesis.it/?p=nethserver-devbox.git;a=blob;f=root/usr/share/nethesis/nethserver-devbox/iso/ks/ks-local.cfg;h=b7dca4098aa29ad3d96572c936ef52d1c21de01e;hb=master
--
Giacomo Sanchietti
Charlie Brady
2013-01-16 17:16:47 UTC
Permalink
Post by Charlie Brady
the rationale for the event queue, etc.
The basic idea is to push into a queue any event signaled by RPM scriptlets
and execute them all only at the end of the yum transaction.
Using the queue we can assure all actions in the %post sections are executed
in the right order respecting dependencies between NethServer modules.
https://dev.nethesis.it/projects/nethserver/wiki/NethServer_yum_plugin
IIUC, when you convert CentOS to nethserver using

http://nethserver.nethesis.it/index.php?id=download_yum-installation

then the Nethserver yum plugin will not be installed, and will not queue
these actions during installation. Have I understood correctly?
Davide Principi
2013-01-16 17:28:50 UTC
Permalink
Post by Charlie Brady
IIUC, when you convert CentOS to nethserver using
http://nethserver.nethesis.it/index.php?id=download_yum-installation
then the Nethserver yum plugin will not be installed, and will not queue
these actions during installation. Have I understood correctly?
Yum plugin is not installed, right, but nethserver-base does the job.
The same applies during media install. See nethserver-base %post
scriptlet.
--
Davide Principi

Nethesis srl - Pesaro (Italy)
Stefano Zamboni
2013-01-17 17:42:24 UTC
Permalink
Post by Giacomo Sanchietti
I can't reproduce the problem. Can you try again? Then tell me date
and time and I will check the logs.
tried right now, same problem, nethserver installed and configured with
default dns

see attached picture

Ciao

Stefano
Stefano Zamboni
2013-01-17 17:44:14 UTC
Permalink
Post by Stefano Zamboni
Post by Giacomo Sanchietti
I can't reproduce the problem. Can you try again? Then tell me date
and time and I will check the logs.
tried right now, same problem, nethserver installed and configured
with default dns
see attached picture
Ciao
Stefano
doh, I forgot this list does not accept binaries..

I'll send you a PM..

S.
Hsing-Foo Wang
2013-01-13 07:35:14 UTC
Permalink
Just checking to see if I understand correctly. Do you mean...

Contribs.org is the SME Server workshop for development, maintenance,
contribs, documentation, ml-lists and discussions. The dedication , time,
resources, donations, developers, community should be dedicated to that
workshop. Any initiative is welcomed, but should take place in our
Contribs.org workshop in order to preserve, consolidate and grow all of the
above?
The question we need to ask ourselves: is SME coming to an end, warranting
throwing away so many years of hard work from so many distinguished
developers, some of the best documentation and contribs around, and above
all a thriving community? Our community.
I think not.
chris
Post by Ian Wells
I would be happy to work on any sustainable continuation of SME Server,
which may be NethServer.
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Ian
______________________________**_________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/**mailman/public/devinfo/<http://lists.contribs.org/mailman/public/devinfo/>
______________________________**_________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/**mailman/public/devinfo/<http://lists.contribs.org/mailman/public/devinfo/>
Giacomo Sanchietti
2013-01-14 17:25:00 UTC
Permalink
Post by Ian Wells
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Actually NethServer is a CentOS with a software layer on top of it, so
we thought this wouldn't be a problem

But, I don't know anything about name and logo usage issue. I will ask
to the CentOS developers.
--
Giacomo Sanchietti
Stefano Zamboni
2013-01-14 18:54:36 UTC
Permalink
Post by Giacomo Sanchietti
Post by Ian Wells
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Actually NethServer is a CentOS with a software layer on top of it, so
we thought this wouldn't be a problem
is it so because it's in alpha stage or is it so "by design"?

S.
Charlie Brady
2013-01-10 03:45:18 UTC
Permalink
Post by Charlie Brady
http://me.veekun.com/blog/2012/04/09/php-a-fractal-of-bad-design/
And also:

http://www.codinghorror.com/blog/2012/06/the-php-singularity.html
Post by Charlie Brady
Whatever you are doing in PHP can be done more securely and more reliably
in perl...
Who wants to help me prove that?
Jean-Philippe PIALASSE
2013-01-10 06:13:51 UTC
Permalink
Hello,

Filippo has point one of the big problem that SME currently has : the
manager interface is not really what we can call user friendly regarding
what we can see with clear OS or even Synology
(http://www.synology.com/us/). Honestly a system with Synology could
really do an excellent job with a really good admin interface.

For who did not try a Synology NAS , you should see their open source
project part : http://sourceforge.net/projects/dsgpl/


The problem now is not for me Perl or PHP as they are server side
language, it is more implementing client side part in order to give a
better user experience. Let's start talking about:
- AJAX
- JQuery
- HTML5
- CSS

Let's also talk about the ability to insert plugins in the main
configurations page instead to create a page per contrib. As an example
, I would rather add a new text box in the ibay page to insert a new
property than creating a new page for this. I would rather to add some
white listing capability in the email setting page rather to have a wbl
contrib page.


I guess that Filippo and his team has chosen PHP because more developers
are familiar with it than Perl. Which could help to see more
contributors. But on the other side, as the heart of the machine use
Perl, they will have to master Perl too ...

I am sure we could create something excellent with jQuery, Ajax and Perl
in order to have a real manager for the 2010's
Post by Charlie Brady
Post by Charlie Brady
http://me.veekun.com/blog/2012/04/09/php-a-fractal-of-bad-design/
http://www.codinghorror.com/blog/2012/06/the-php-singularity.html
Post by Charlie Brady
Whatever you are doing in PHP can be done more securely and more reliably
in perl...
Who wants to help me prove that?
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Davide Principi
2013-01-10 16:52:25 UTC
Permalink
Post by Jean-Philippe PIALASSE
The problem now is not for me Perl or PHP as they are server side
language, it is more implementing client side part in order to give a
- AJAX
- JQuery
- HTML5
- CSS
We totally agree: you can find this stuff in NethServer UI (see
http://nethserver.nethesis.it/index.php?id=modules_web-gui)
Post by Jean-Philippe PIALASSE
Let's also talk about the ability to insert plugins in the main
configurations page instead to create a page per contrib. As an example
, I would rather add a new text box in the ibay page to insert a new
property than creating a new page for this. I would rather to add some
white listing capability in the email setting page rather to have a wbl
contrib page.
Same as above. This is a key feature of NethServer UI. For example the
nethserver-httpd rpm -- when installed -- adds a UI plugin to "Shared
Folders" (ibays) UI module, showing a tab with HTTP access options.
Post by Jean-Philippe PIALASSE
I guess that Filippo and his team has chosen PHP because more developers
are familiar with it than Perl. Which could help to see more
contributors. But on the other side, as the heart of the machine use
Perl, they will have to master Perl too ...
We are providing a UI framework (Nethgui) aimed at hiding some of the
(complex) tasks of web UI technologies and programming. Yes, we have to
improve documentation for developers, but we want to make the PHP
programming as simple as possible.
Post by Jean-Philippe PIALASSE
I am sure we could create something excellent with jQuery, Ajax and Perl
in order to have a real manager for the 2010's
The web UI is served by nethserver-httpd-admin package. It can be
replaced by any other system that sets DB and signals events properly!
--
Davide Principi <***@nethesis.it>
Nethesis srl
Filippo Carletti
2013-01-10 08:51:06 UTC
Permalink
Post by Charlie Brady
Post by Filippo Carletti
We would like to attract developers and sysadmins
...
Post by Charlie Brady
Post by Filippo Carletti
server-manager interface, in php.
I think that choice was very unfortunate.
While I agree with you on php "shortcomings", it's a well known
language. We wanted something known by many, so that a casual
developer could be immediately productive when trying to write a
contrib interface (the "hardest" part of writing a contrib now is the
panel, look at how many db properties are not exposed on the
server-manager).

We are open to suggestions, that's why I started this thread.
We'll evaluate any suggestion. I can't speak for frameworks, I'll ask
my team mates to evaluate them.

--
Ciao,
Filippo
Charlie Brady
2013-01-09 18:49:55 UTC
Permalink
Post by Filippo Carletti
http://nethserver.nethesis.it
Maybe I am missing it, but I don't see hardware requirements listed there.
I tried a test install on a small VM, but it fails "Not enough space for
LVM requests". During development at least, I would recommend you support
installation on small footprint systems.

Thanks
Hsing-Foo Wang
2013-01-09 18:53:49 UTC
Permalink
FYI, I've installed it on a Parrallel's VM (no tools installed):
- 1 CPU
- 1Gb RAM
- 15Gb HD
Post by Charlie Brady
Maybe I am missing it, but I don't see hardware requirements listed there.
I tried a test install on a small VM, but it fails "Not enough space for
LVM requests". During development at least, I would recommend you support
installation on small footprint systems.
Thanks
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Michel-André
2013-01-09 19:28:05 UTC
Permalink
Hi all,

Same thing, I went up to 8GB then 16GB and it was OK.

Michel-André
*****************************************************
Post by Charlie Brady
Post by Filippo Carletti
http://nethserver.nethesis.it
Maybe I am missing it, but I don't see hardware requirements listed there.
I tried a test install on a small VM, but it fails "Not enough space for
LVM requests". During development at least, I would recommend you support
installation on small footprint systems.
Thanks
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
Charlie Brady
2013-01-09 20:14:42 UTC
Permalink
Post by Michel-André
Hi all,
Same thing, I went up to 8GB then 16GB and it was OK.
I do not have 16GB free on my development laptop, so I will wait and see
if a new version solves this issue.
Hsing-Foo Wang
2013-01-10 00:04:06 UTC
Permalink
Post by Charlie Brady
I do not have 16GB free on my development laptop, so I will wait and see
if a new version solves this issue.
Since nobody knows the state of donations, or the spending of it, please
tell me you brand/model and I privately will happily sponsor a new HD for
your laptop with the largest GB available. You have my email address.

Seems to me that a 'waiting' attitude due to such a small issue does not
help or challenge us, as in community incl. devs, at all. Innovation and
progress, under the cover of the long standing security and
carefulness philosophy of SME Server, at this stage of SME Server seem more
important. IMHO.

I feel we really need to re-work our internal/external marketing without
disrespecting anybody!

Regards,
-HF
Charlie Brady
2013-01-10 02:47:18 UTC
Permalink
Post by Charlie Brady
Maybe I am missing it, but I don't see hardware requirements listed there.
I tried a test install on a small VM, but it fails "Not enough space for
LVM requests".
I discovered that I could install a VM with 6GB disk by first installing
CentOS 6.3 minimal iso (x86_64), and then mounting the Nethesis iso, then
doing:

cd /media
rpm -Uhv extras/nethserver/*.rpm
/sbin/chkconfig fcoe off
/sbin/chkconfig iscsi off
/sbin/chkconfig iscsid off
/sbin/chkconfig netfs off
/sbin/chkconfig nfslock off
touch /var/spool/first-boot
reboot

The console configuration wizard comes up after restart, and the web
interface is then available on https://server.ip.address:980/.
Filippo Carletti
2013-01-10 08:15:45 UTC
Permalink
Post by Charlie Brady
I discovered that I could install a VM with 6GB disk by first installing
CentOS 6.3 minimal iso (x86_64), and then mounting the Nethesis iso, then
After CentOS 6.3 minimal install, please follow instructions here:
http://nethserver.nethesis.it/index.php?id=download_yum-installation


--
Ciao,
Filippo
Charlie Brady
2013-01-10 13:32:03 UTC
Permalink
Post by Filippo Carletti
Post by Charlie Brady
I discovered that I could install a VM with 6GB disk by first installing
CentOS 6.3 minimal iso (x86_64), and then mounting the Nethesis iso, then
http://nethserver.nethesis.it/index.php?id=download_yum-installation
AFAICT, what I have already done accomplishes exactly the same set of
tasks. Is there something I have missed?
Davide Principi
2013-01-10 14:33:46 UTC
Permalink
Post by Charlie Brady
Post by Filippo Carletti
http://nethserver.nethesis.it/index.php?id=download_yum-installation
cd /media
rpm -Uhv extras/nethserver/*.rpm
/sbin/chkconfig fcoe off
/sbin/chkconfig iscsi off
/sbin/chkconfig iscsid off
/sbin/chkconfig netfs off
/sbin/chkconfig nfslock off
touch /var/spool/first-boot
reboot
AFAICT, what I have already done accomplishes exactly the same set of
tasks. Is there something I have missed?
That resembles the kickstart command list that disables some default
services coming from CentOS minimal. It does not hurt.
--
Davide Principi <***@nethesis.it>
Nethesis srl
Stefano Zamboni
2013-01-09 19:17:05 UTC
Permalink
Post by Filippo Carletti
Smeserver users and developers,
Hi Filippo :-)
Post by Filippo Carletti
during the last months, we (Nethesis) tried to build what we think
NethServer
- simple configuration
- quick trouble shooting
- fast development of new features
I remember a chat about it, many months ago :-)
I'm happy to see you are going on..
Post by Filippo Carletti
- a new dynamic, streamlined user-interface
- standard components, tested widespread configurations
- minimal differences with upstream
clear enough
Post by Filippo Carletti
We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
I agree with you, again
Post by Filippo Carletti
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php. We rebuilt the mail server using
postfix, amavis and dovecot.
Since nethserver could be (let me tell) the future of SME, is there any
way to run current contribs (not bound to qmail, for example) on
nethserver?

are db files the same? I mean, same structure?
Post by Filippo Carletti
We are still working on many features and writing documentation for
contribs developers, but we hope to have reached a point were an alpha
release is good enough to have your opinion.
This alpha release is well tested but not ready for production environments.
NethServer is released under the GPL, development is open, developers
are welcome.
http://nethserver.nethesis.it
I'm trying to download.. I'm not the only one I think, because it's
sloooooooooooow.. any plan to share it via torrent?


finally, I add just my 2c: I think SME is a great product but I
sincerely can't see a future..
I hope that every SME experienced developer will adhere to your
project..
yes, Charlie, I'm referring to you :-)
out there there are plenty linux distro aimed to small/medium
enterprises.. no one is, IMHO, complete like SME and NO ONE has a wiki
and a support like SME..
we all know that security is the first goal.. and we all know that php
is sometime very insecure.. but, as long as server-manager app(s) is
locked to lan only, its code checked and verified, we should not be
afraid

I will try nethserver asap

Filippo, good job..
<italian only>
salutami tutti i ragazzi di Nethesis, prima o poi verrò a trovarvi di
nuovo :-)
</italian only>

at ALL guys reading here: this COULD be the last train.. I hope we don't
loose it

<final disclaimer>
I'm not involved, in any way, with Nethesis.
I met them 2 years ago, it was a beautiful experience
</final disclaimer>

Ciao

Stefano
Hsing-Foo Wang
2013-01-09 19:23:54 UTC
Permalink
Post by Stefano Zamboni
<final disclaimer>
I'm not involved, in any way, with Nethesis.
I met them 2 years ago, it was a beautiful experience
</final disclaimer>
How did we miss that invite?......
Stefano Zamboni
2013-01-09 19:27:04 UTC
Permalink
Post by Stefano Zamboni
<final disclaimer>
I'm not involved, in any way, with Nethesis.
I met them 2 years ago, it was a beautiful experience
</final disclaimer>
How did we miss that invite?......
I'm not sure I fully understood your message but..

It was at a party, I was on holydays near to them :-)

ciao

S.
Charlie Brady
2013-01-09 23:41:01 UTC
Permalink
Post by Stefano Zamboni
we all know that security is the first goal.. and we all know that php
is sometime very insecure.. but, as long as server-manager app(s) is
locked to lan only, its code checked and verified,
Do you know anyone capable of checking and verifying the security of
non-trival PHP applications? Even if you do, that will not be sufficient,
since there have been many security flaws in the PHP interpreter itself.
Post by Stefano Zamboni
we should not be afraid
Sorry, but I don't agree.
Filippo Carletti
2013-01-10 08:27:00 UTC
Permalink
Post by Stefano Zamboni
Since nethserver could be (let me tell) the future of SME, is there any
way to run current contribs (not bound to qmail, for example) on
nethserver?
If the contrib doesn't have an interface it could work. We're writing
a guide on how to port contribs to NethServer.
Post by Stefano Zamboni
are db files the same? I mean, same structure?
Yes, the db format is unchanged.
Post by Stefano Zamboni
I'm trying to download.. I'm not the only one I think, because it's
sloooooooooooow.. any plan to share it via torrent?
It's served from akamai cdn, it should be as fast as your link.
Working on a torrent, thanks for the idea.


--
Ciao,
Filippo
Huib de Visser
2013-01-13 10:59:37 UTC
Permalink
Why not refocus development here at contribs.org with all it's facilities, community and where the longstanding devs with specific and intimate knowledge may provide guidance with the limited time they have, and give Nethesis the necessary access to (co-)develop.

In return give Nethesis acknowledgement of their corporate backing (maybe like: logo on the contribs.org website, credit on a start page when accessing the server-manager interface, credit it the code, ...)

I do like some new fancyness in the WebGUI, but when in comes to servers i'd say safety first and try to implement new webgui features in perl with some sort of framework that is undoubtedly out there.



Just my2cents,
Huib.


-----Original message-----
To: Contribs - DevInfo <***@contribs.org>;
From: Filippo Carletti <***@gmail.com>
Sent: Wed 09-01-2013 12:30
Subject: [devinfo] NethServer: something new to play with
Post by Filippo Carletti
Smeserver users and developers,
during the last months, we (Nethesis) tried to build what we think
NethServer
- simple configuration
- quick trouble shooting
- fast development of new features
- a new dynamic, streamlined user-interface
- standard components, tested widespread configurations
- minimal differences with upstream
We would like to attract developers and sysadmins, so we hope that
having a common, well known environment would be a good start.
We use and like Smeserver, we think that templates and events are a
great idea, so we retained them, along with the plain text
configuration databases in NethServer. We built from scratch the new
server-manager interface, in php. We rebuilt the mail server using
postfix, amavis and dovecot.
We are still working on many features and writing documentation for
contribs developers, but we hope to have reached a point were an alpha
release is good enough to have your opinion.
This alpha release is well tested but not ready for production environments.
NethServer is released under the GPL, development is open, developers
are welcome.
http://nethserver.nethesis.it
https://groups.google.com/d/forum/nethserver
Thank you.
--
Ciao,
Filippo
_______________________________________________
Server Development Discussion
Searchable archive at http://lists.contribs.org/mailman/public/devinfo/
John Crisp
2013-01-13 13:40:01 UTC
Permalink
Post by Filippo Carletti
Smeserver users and developers,
Having read the comments on this thread I have to say I feel somewhat
disappointed, and have an awful sense of foreboding about the whole thing.

I have used the contribs distro for a number of years and think it fits
a fantastic niche in the market.

Right now though I feel somewhat railroaded by current developments, and
after only a couple of days I can see that shortly I am going to have to
make some decisions about my future server software, and it has made me
start to look at a wider playing field. How many other users will do the
same ?

In the first instance I commend Filippo for his work and take nothing
away from that. Obviously anyone is entitled to fork (and that is
exactly what this is) and develop as they see fit. But the actions are
instantly dividing the contribs community. And this is no good for
either distro.

Surely it would have been far better to have started a conversation on
the 'discussion' lists before development work began to sound out the
community on its future, which is where this discussion really belongs.
After all, it's about the future of contribs as much as anything. It
would be far better to get the majority on board, which is not
necessarily the case right now.

I now feel that there have been commitments made for which there is no
return, and I get the sinking feeling in the pit of my stomach about the
whole thing. Kind of like watching the unsinkable ship hitting an
iceberg. You just know what the outcome is going to be be, even if it
takes a while.

I don't dispute that contribs needs a future beyond V8. Maybe it will be
NethServer. With better discussion, maybe there could have been a better
distro.

Unfortunately, I just feel that the whole distro has been thrown into a
crisis it really doesn't need. That isn't good for attracting new users.

And for that I feel very sad.

I will open a thread on 'Discussions' which I think is a better list for
comments on the future of contribs, and perhaps the contribs community
can give some serious thoughts on the way ahead.

B. Rgds
John

PS like others, though interested in developments in NethServer, I won't
be joining a Google Group......
Pascal Schirrmann
2013-01-13 15:32:41 UTC
Permalink
Post by John Crisp
Post by Filippo Carletti
Smeserver users and developers,
Having read the comments on this thread I have to say I feel somewhat
disappointed, and have an awful sense of foreboding about the whole thing.
Hi, everybody,

Long time since I wrote on this list...

I had a quick look at netserver, and, from my point of vue, itis not
really a technical fork of SME, but a spiritual fork :
I think that the idea behind is to have a 'core CentOS', + a
configuration system relying on templates.
Obviously, this tend to a big lot of changes, and not a lot of existing
contribs will work in that system.
OTOH, this is probably one of the best ideas to have (after the first
BIG step) a fairly easy to upgrade system.

So, basically, this will split the people in two groups, and, until one
group wins, it will be war.

I would prefer that, before the war, a "constructive" discussion is
set-up about the technical choice.

My 2 cents,
Pascal
Dave Liquorice
2013-01-14 22:25:44 UTC
Permalink
Post by Giacomo Sanchietti
Post by Ian Wells
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Actually NethServer is a CentOS with a software layer on top of it, so
we thought this wouldn't be a problem
But, I don't know anything about name and logo usage issue. I will ask
to the CentOS developers.
What are you "selling" CentOS or the fruits of your hard labours,
NethServer?

Irrespective of any rights issues with logos and names do you really want to
be hiding your identity in the market place?
--
Cheers
Dave.
chris burnat
2013-01-14 22:46:17 UTC
Permalink
Post by Dave Liquorice
Post by Giacomo Sanchietti
Post by Ian Wells
First impression was that you need to make urgent changes to respect the
branding - i.e. CentOS name and artwork are visible.
Actually NethServer is a CentOS with a software layer on top of it, so
we thought this wouldn't be a problem
But, I don't know anything about name and logo usage issue. I will ask
to the CentOS developers.
What are you "selling" CentOS or the fruits of your hard labours,
NethServer?
Irrespective of any rights issues with logos and names do you really want to
be hiding your identity in the market place?
Dave, This kind of comment does not contribute in any shape of form to
the important dialogue in progress with the very civilised crew at
NethServer about the future of the distro. Please show restraint and
refrain from further questionable posts.
Thanks
chris
Ray Mitchell
2013-01-21 16:57:12 UTC
Permalink
Filippo

Your NethServer project is to be commended. The design approach is
innovative. The modular approach will allow users & developers to not use or
remove or rewrite modules.

I think in order to gain general support from the existing SME Server
community, it would be wise to listen to suggestions made by Charlie Brady &
Gordon Rowell, they are the SME experts. Community members respect their
opinions & technical understanding, so their suggestions on technical merit
should be taken seriously. After all this is your first attempt at an SME
Server release, whereas Charlie & Gordon have 5 or 6 major versions worth of
experience (since sme 3.x). They know all the pitfalls & traps & I'm sure
they keep well informed on current trends.

You do seem to be "listening" & are "willing to take advice on board", so
that is good.
Whether that will translate to major changes to your Alpha release code is
yet to be seen, eg in particaulr re the PHP server manager.
I go with Charlie & the Mojolicious approach.

There is still much work to be done to (hopefully) incorporate community
recommendations & requests, and to engage the community to support your
project. If you can incorporate their advice & recommendations, it is
certainly possible your project may become the "next" SME Server
replacement, based on CentOS 6.x. Obviously much of that "advice &
recommendations" is yet to be forthcoming.
I think people are still basically seeing & learning what your Alpha release
is all about, and also assessing your general attitude towards the
community. I'm sure most of us are wondering "how & what" you intend to
blend into your efforts with the community, whether you will remain a
standalone group "doing your own thing" or whether you bring more community
support, involvment, recommendations & infrastructure usage into the (new)
NethServer world. It's probably a BIG question for now, to which answers do
not yet exist. You too are seeing what reaction you are getting from the
community.

If you can come up with a good replacement that the majority of the
community is happy with, especially re technical "soundness", then I could
see it being fully embraced.

You probably also need to more fully consider how to engage or utilise the
existing SME Server community infrastructure, so the project appears
community based. If the community does not have a sense of "ownership" &
perhaps also "control", then those who do the work behind the scenes may not
want to support your server project.

I think it has been mentioned, but a migration path from SME Server 8.0 to
NethServer will surely be a necessity, either by an upgrade plus migration
routine, or a backup & restore approach (I'd assume also with a suitable
migration routine involved).
I personally would hate to have to rejoin Windows workstations/users to the
SME Server Domain Controller & setup Domain user profiles again, not to
mention a whole lot of other data & configuration issues etc. This need is
some time away yet, but it has always been a major feature of SME server.

Hope these comments help.

Regards
Ray
Ray Mitchell
2013-01-21 17:00:01 UTC
Permalink
Filippo

Re NethServer Backup:
Rather than just reproducing existing backup functionality, perhaps look at
smeserver-rdiff-backup-0.0.1-19.el4.sme.noarch.rpm
as your backup method.
It's CLI only at present, but I'm sure you could create a GUI panel.
Some notes here
http://forums.contribs.org/index.php/topic,35525.msg155709.html#msg155709

Old it may be, but AFAIK it's still being used as the "preferred backup
method of choice" by many professional SME server support personnel, who
tend to prefer the CLI interface (I think for similar reasons that Affa is
only CLI).

Regards
Ray

Continue reading on narkive:
Loading...