How many of you have switched/migrated from SME Server?

Hi Stefano

I seem to remember having a conversation with you some time ago on the SME community pages regarding Postgresql. There was a howto for version 7+ that didn’t translate to version 8+. The problem was the repos (I think…) but whatever, all I could manage to install was version 8.2 and I wanted 9.1. It would seem 8.2 was the default for some reason . The bottom line is, with all that with Ubuntu server it was so simple, and there is the key. The good thing about the exercise was that I did stop to think about implications of having data on a network server, as opposed to a dedicated and protected database server.

Cheers

Hi Alessio

Yes I saw that…but I much prefer to have my data on a separate dedicated server. But it is good to know that Postgresql can be installed if required. Good work.

Cheers

I’ve been running SME Server for a number of years (installed the first one version 7 one in 2008 if I remember correctly). Currently I run admin on 12 SME boxes. Most of them are mail/internet gateways in local small businesses.

Nethserver is appealing due to its more modular design and due to its use of dnsmasq as opposed to the tinydns setup that SME uses.

I’ve hit a few problems in my first, trial migration. First, after reading the migration guide, I didn’t realise that the migration did not import data from the SME backup, only settings. It also seems to have been quite selective in some areas - it imported things like /root/.ssh, but failed to pick up /root/.synbak/tar. Not a major problem, just not what I was expecting.

I performed the Nethserver install to an HP 8000 small form factor machine. I used the unattended install so that the new server would come up with DHCP enabled so that it would get an IP in my local network and I can then drop it into the target network before setting a static IP for it. Worked fine.

I opened a current complete backup from the SME server, expanded the archive to /var/lib/migration, then ran the migration. I’ve not yet copied the data to the test server - I’ll have to get a current copy of the data when I put the server into the target network.

Questions. Is there an equivalent to the “signal-event post-upgrade; signal-event reboot” that SME uses to expand all templates?

On SME, I do a lot of remote admin, and find it very useful to be able to use elinks to access the server manager during an SSH session. Is it possible to use a text-mode browser with the Nethserver server manger? If so, what url do I need to use to access it?

One of the feature in SME that I’ve found useful in the past is the ability to limit server manager access to specific IPs. Is it possible to do something like that with Nethserver?

I’m also wondering if there is a guide to setting firewall rules? I need NFS support on the target server, and so far, I’ve not managed to get the firewall set up to allow NFS access from machines in the local net.

:smile: I’m sure I’ll have more questions once I get a bit further with the migration…

No, the migration imports everything: settings AND data.
Obviously, data has to be present in the backup. :smile:

AFAIK, the whole /root directory is copied (I did tens of migrations).

I’d say NO, because NS don’t need such a big hammer. But to expand all templates and restart the relevant services (no reboot, a business server shouldn’t be rebooted) you can use:

 /etc/e-smith/events/actions/system-adjust

Yes:

elinks https://localhost:980

Access to all network services can be limited through the Network services menu.

2 Likes

I got the error

Unable to retrieve https://localhost:980/: 
                  SSL error

I think that elinks needs a “valid” (read commercial) certificate.
Please try something like:
elinks -eval ‘set connection.ssl.cert_verify = 0’ https://localhost:980/

1 Like

Your solution is good…But the display is broken, you should use a vpn solution for accessing to the Server Manager

Ehi Paul thanks for your sharings and for having accepted my invite. Please don’t give up and keep us updated on your migrations :+1:

On 12/01/16 22:52, Filippo Carletti wrote:

filippo_carlettiDev Team

January 12paul_marwick:

I didn't realise that the migration did not import data from the SME backup, only settings.

No, the migration imports everything: settings AND data.

              Obviously, data has to be present in the backup. <img moz-do-not-send="true" src="imap://paul%2Emarwick@imap.gmail.com:993/fetch%3EUID%3E/%5BGmail%5D/Drafts%3E1050?v=0" title=":smile:" alt=":smile:" style="max-width:100%;" height="20" width="20"></p>
:) My error - I had forgotten that NS uses different locations for
mail and iBays. The data was there...

paul_marwick:

it imported things like /root/.ssh, but failed to pick up /root/.synbak/tar

AFAIK, the whole /root directory is copied (I did tens of migrations).

On my second attempt it was, but for some reason some things weren't
on the first attempt.
paul_marwick:

Is there an equivalent to the "signal-event post-upgrade; signal-event reboot" that SME uses to expand all templates?

I'd say NO, because NS don't need such a big hammer. But to expand all templates and restart the relevant services (no reboot, a business server shouldn't be rebooted) you can use:

Thanks. The reboot sequence that SME always seems to need is one o
fmy least favourite features I must admit. I'll keep that one in
mind.
 /etc/e-smith/events/actions/system-adjust
paul_marwick:

Is it possible to use a text-mode browser with the Nethserver server manger?

Yes:

elinks https://localhost:980

I’ve not managed to get that to work. Elinks comes up with SSL
errors. Lynx gets a bit further, but still fails to load the
server manager.


paul_marwick:

ability to limit server manager access to specific IPs

Access to all network services can be limited through the Network services menu.

I’ve not looked closely at that yet.
I ended up trying again, twice. I ended up with errors both times, so I must be doing something wrong. I’ve captured the section of /var/log/messages just before the error starts:

Jan 13 16:17:57 goldie esmith::event[28751]: Migrating existing
database fetchmail


Jan 13 16:17:57 goldie esmith::event[28751]: Migrating existing
database dhcp

A post was split to a new topic: Locked out of mysql

After installing elinks with “yum install elinks” I tried the command but got an SSL-Error…

Is there another way to access the server-manager via ssh ?

Occasionally I use SSH to forward a local port to the remote host. IIRC…

ssh -L localhost:9800:localhost:980 -p 22 sshuser@fqdn

then on the web-browser go to https://localhost:9800 to access the server-manager.

2 Likes

the links2 CLI browser seems to be working (at least on NethServer 7):

yum install links
links2 https://localhost:980
2 Likes

OK

yum install links
links2 https://localhost:980

This works for me… thanks !

great! :raised_hands::clap:
this saved me after a problem with openvpn… please write it on wiki
tnx!

2 Likes

Done

3 Likes

Quoting this from another discussion

I’d like to write together a doc on our wiki where collect all advice about migration from SME Server

@kieronrob @Bluelake @dnutan @paul_marwick @fausp @macwunder @flatspin could you help?

@Ale could you lead us? This discussion and that doc might be a good starting point Migration from NethService/SME Server — NethServer 7 Final
Don’t worry about your English we have many natives here that can correct our writings

Tagging also @Crazyusb @xmechanic @haliparotin @bjhankins @Erwis @bernhard @drifting@telekomiker @kieronrob @i_combo @panzerphreak @jfranco

I tried it on NS6.9 and it worked.

@alefattorini

I support about 20 Clients with their Network and IT System.

I’ve been using Linux Servers from the beginning - even though others called me a MS-Guru.
Started with SuSE Linux, still have clients using GroupWise / Linux but for most cases that’s too complicated. So tested out a few other distros and settled on SME5/SME6 at the time.

I also might add that I’m a VMWare user since 1999 (Yes, VMWare 1.0 beta!) but recently have moved to ProxMox. While I still consider VMWare a premium grade Virtualisation Platform, I don’t like the idea of needing a MS Server to administrate a Linux box, nor several more if i really want to use ESXi.

Same goes now for SME Linux. While it’s absolutely stable and rock solid it’s still stuck on it’s EL6 base, including very old PHP. And LetsEncrypt was the game changer. While I found the name amusing - considering the old PHP background - no, I just wasn’t going to install a package called dehydrate to be able to use LetsEncrypt.

Using OwnCloud, later NextCloud on SME was a PITA, and it never really worked well, just barely, again because of old PHP.

SME also wanted too many reboots. I mean, updating something like jpg-libs or such doesn’t really need the whole server to reboot!

I’ve actually only done one migration from SME, the rest was manual transfering. That transfer went quite well. The others had too many legacy stuff I didn’t want migrated!

Nethserver has the same base, but runs on newer Centos EL7.
OK, NextCloud needs a PHPSCL environment like SME, but with less headache. It works fairly well.

And yes, I’m one of the Mac users here. I can handle Windows and Linux extremly well, but my personal boxes are all mac. And one PiTop (Raspberry) Notebook. My Clients use Windows more than I do, but they also use Mac and Linux.

Besides the 7.4 s*it served from upstream, NethServers been a great server!
Let’s keep it that way!

Best regards from Switzerland!
Andy

1 Like

filippo_carletti
Dev Team

            January 12

paul_marwick:

                  I didn't realise

that the migration did not import data from the
SME backup, only settings.

              No, the migration

imports everything: settings AND data.

              Obviously, data has to be present in the backup.
:) My error - I had forgotten that NS uses different locations for

mail and iBays. The data was there…

<img moz-do-not-send="true" alt="" src="https://avatars.discourse.org/letter/p/a9adbd/40.png" style="margin-right: 5px; width:20px; height:20px;
  max-width:100%;" height="20" width="20">paul_marwick:
                  it imported

things like /root/.ssh, but failed to pick up
/root/.synbak/tar

              AFAIK, the whole

/root directory is copied (I did tens of migrations).

On my second attempt it was, but for some reason some things weren't

on the first attempt.

paul_marwick:

                  Is there an

equivalent to the “signal-event post-upgrade;
signal-event reboot” that SME uses to expand all
templates?

              I'd say NO, because

NS don’t need such a big hammer. But to expand all
templates and restart the relevant services (no
reboot, a business server shouldn’t be rebooted) you
can use:

Thanks. The reboot sequence that SME always seems to need is one o

fmy least favourite features I must admit. I’ll keep that one in
mind.

>  /etc/e-smith/events/actions/system-adjust

paul_marwick:

                  Is it possible

to use a text-mode browser with the Nethserver
server manger?

Yes:

> elinks [https://localhost:980](https://localhost:980)
`      I've not managed to get that to work. Elinks comes up with SSL

errors. Lynx gets a bit further, but still fails to load the
server manager.

`

paul_marwick:

                  ability to limit

server manager access to specific IPs

              Access to all

network services can be limited through the Network
services menu.

I've not looked closely at that yet.

I ended up trying again, twice. I ended up with errors both times,

so I must be doing something wrong. I’ve captured the section of
/var/log/messages just before the error starts:

Jan 13 16:17:57 goldie esmith::event[28751]: Migrating existing

database fetchmail

Jan 13 16:17:57 goldie esmith::event[28751]: Migrating existing

database dhcp

Jan 13 16:17:57 goldie esmith::event[28751]: Action:

/etc/e-smith/events/nethserver-mysql-update/S00initialize-default-databases
SUCCESS [0.240041]

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]: NOTE: RUNNING ALL PARTS

OF THIS SCRIPT IS RECOMMENDED FOR ALL MySQL

Jan 13 16:17:58 goldie esmith::event[28751]:       SERVERS IN

PRODUCTION USE! PLEASE READ EACH STEP CAREFULLY!

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]: In order to log into

MySQL to secure it, we’ll need the current

Jan 13 16:17:58 goldie esmith::event[28751]: password for the root

user. If you’ve just installed MySQL, and

Jan 13 16:17:58 goldie esmith::event[28751]: you haven't set the

root password yet, the password will be blank,

Jan 13 16:17:58 goldie esmith::event[28751]: so you should just

press enter here.

Jan 13 16:17:58 goldie esmith::event[28751]:

Jan 13 16:17:58 goldie esmith::event[28751]: Enter current password

for root (enter for none):

Jan 13 16:17:58 goldie esmith::event[28751]: ERROR 1045 (28000):

Access denied for user ‘root’@‘localhost’ (using password: NO)

Jan 13 16:17:58 goldie esmith::event[28751]: Enter current password

for root (enter for none):

Jan 13 16:17:58 goldie esmith::event[28751]: ERROR 1045 (28000):

Access denied for user ‘root’@‘localhost’ (using password: NO)

Jan 13 16:17:58 goldie esmith::event[28751]: Enter current password

for root (enter for none):

Jan 13 16:17:58 goldie esmith::event[28751]: ERROR 1045 (28000):

Access denied for user ‘root’@‘localhost’ (using password: NO)

In the first case, I'd set up the correct domain information and the

name of the server that the backup is derived from. I’d also updated
the NS installation and added the necessary packages to cover the
functions that I needed. In the second, I’d not set the domain to
what it should be eventually, but the error was the same in each
case.

I'll have another look in a couple of days - too many jobs to do for

the next few days.

Paul