My first longtime NS7 Client migrated to NS8 RC1 ! (Yes, I trust the Devs!)

Hi Ereryone reading this!

I am a longtime NethServer user, migrating at around 2015 from SME Server to NethServer for all of my roughly 30 clients. At around the same time, I quit using VMWare ESXi and migrated to using Proxmox as my chosen Hypervisor.

I’ll say today: I never regretted both moves!!!

The past Weekend, I migrated a client of mine, a doctors practice, to NS8 RC1. No, not even released!!!
But truth is, I trust the people in the Dev Team so much that I migrate before it’s even released!

And I must say, I wasn’t disapointed!

The system is working from day one to the pleasure of all involved.

And: To all not aware of it, a doctors practice often involves more than the usual plethora of devices like servers, workstations, storage (NAS), Tablets and Smartphones… How about X-Ray machines, Ultrasonic Imageers all trying to store their DICOM images to some AD integrated Storage…

Even though the AD changed it’s IP, all clients, even the “devices” found the needed information and were able to access storage, servers, what ever!

→ The AD in NS8 seems to work very well!

Expose more features in the GUI, please!

My Kudos to all Devs involved, but also to the users of this platform / Forum, contributing in testing and reporting found bugs!

Great Work, Everyone!

My 2 cents
Andy

12 Likes

If AD changed the DC IP, how do clients find the new one?

Did you see this PR for the migration instructions of AD? NethServer 7 migration — NS8 documentation

Do you suggest an alternative approach?

Hi @davidep

The IP change is mandatory in NethServer.

NS7 has an IP different from itself, used as AD (Must be a free IP in the same subnet!).

Using the NS8 migration, at the end of the migration, it will request you choose an IP from a dropdown box for use of the AD. The only IP shown is the IP designated to Node1 of the new Cluster.

This is per SE a different IP for the AD than earlier NS7 used (The AD must be accessible during the migration, so using the same IP is de facto impossible!).

However, I am known also as a “DNS Guru” for over 30 years now, and AD is based on DNS, so it’s nothing I wasn’t aware of before the migration (Or can’t handle!). I also have a DNS server capable of handling CNAMEs, so I created aliases where necessary and it all added up - and worked!

The actual process is called DHCP, but added assistance from DNS fine tuning with Aliases and a defined “switch Over” moment! :slight_smile:

And yes, I am one of those who do actually Reads The Fine Manuals… :slight_smile:
Those pages are still open on my browser since the start of the migration, Friday evening!

Yes, I do have some suggestions for this.

As the target is known (The new NS8), the migration tool should query it first, eg using a LDAP query.
Authentification details are known to the user, so the user can provide these.

This would lower errors from migrations, alleviate false infos and provide a smoother experience.
It’s “tested and queried”, not “assuming”… :slight_smile:

My 2 cents
Andy

2 Likes

Right, it is worth to list it too with the current choices:

Help me to understand. If the migration tool sends an LDAP probe to the future DC IP, it receives no answer because there is nothing running there, not yet. At least until migration is finished. What am I missing?

That would be the Idea. In my case, I was getting the messge AD available on NS8, even though the NS8 was a vanilla cluster, just set up!
And REFUSING the migration or any step to connect!

In present times, that’s called Fake News"!

My 2 cents
Andy

1 Like

Ok, understood! I’m sure it was a false alarm, maybe raised by a bug: I hope to find it, as said in my PVT message.

3 Likes

Did you migrate also shared folders to NS8? How much data did it transfer?

Hi

Yes, I also migrated all shared folders from that NS7 to the new NS8. Volume of data on that NethServer was around 80 GB, one of the smallest as file servers from my clients.
No issues during transfer!

1 Like

What about file ACLs? Are they fine grained, or set at default for the share and its contents?

ACLs are not fine grained, there are basically only 2 levels needed, this is set by groups.
There only 5 employees, all long term, and two doctors, so simple permissions wished by client…
The shares are all working without issues.

As said, I trnsfered a weeks earlier NethServer7, so data for that week was missing.
I retrieved the 4 shares from proxmox as ZIP file (PBS allows file / folder restores directly from the backup, no need to restore a VM.) and used a virtual Windows VM to transfer the extracted ZIPs back into the folders they came from.

Ergo, the server had the correct contents from Friday evening.

1 Like

Got it: you didn’t use the rsync transfer provided by the migration package. I guess you selected the “skip migration” button in the tool UI for the File server part.

No, I actually used it - also as a test, even though I knew I needed to restore data.
→ I still have quite a few NethServer 7 installs to migrate.

I had a fast server, space and time for 80 GB all in all was negliciable for me, so why not?

And f it becomes better in future due to feedback, I’m among the first to provide it.

:slight_smile:

Spill the beans, Andy.
Which modules were installed on NS7?

Hi @pike

The “hot” mexican beans, or the specialized ones increasing flatulance efficiency? :slight_smile:

No problem with that, here’s the screenshot showing all installed stuff:

Zabbix & Guacamole remain installed there, at least until Zabbix is migrateable, or can be transfered to NS8 (Or a VM/Container in worst case…)

As Nextcloud is hardly used here (Most of my clients use nextcloud fairly intensively, but ba far not all), I decided against migrating this very useful App.

Mail is not used locally at all, but rather HIN, a government secure mail much like what Italy has, but only for anyone in the medical / health business. Then its actually mandatory!

Hope this covers your justified curosity!

Some of my other clients have much more complex environments…

This client I deemed an “easy” one, as the major challenges are the AD (And not common devices using / connecting to AD), the need to change the internal / external Domain name. The old owner is pensioned and may want to use that old dmain for family purposes, the new docctor / owner has a new domain which can be used instead.
Only the AD must remain, at least internally, as we would have to reinstall the whole doctors software, which the (expensive) specialists have to handle… This runs on a Windows VM as “Member Server” to run Apps.

My 2 cents
Andy

2 Likes

I find it an exciting read.
If I understood correctly, you simply reinstalled NextCloud after the migration - right?

How did the migration of the vhosts behave? Were the web servers with the databases completely migrated and ready to run? What about the PHP versions? Were you able to update them to the latest version afterward?

Congrats, Marko

Congrats! on a successful migration to NS8 RC1 for your customer. I am curious what Linux distro you decided to use for your underline distro (CentOS/AlmaLinux/RockyLinux or Debian) for the migration?

-SF-

@Shadowfire

Hi Jeff

I did write here that I’ve lost trust in anything “handed down” from RH, Big Blue or any of their Followers.
I don’t like the Term, as a “Follower” can never be a leader!
So I can’t see Rocky, Alma or even Oracle as a leader.
And who would want to “follow” a parrot in real life anyways?

Go Team Debian !!!

  1. It’s been around LONGER than Red-Hat
  2. Hard to be bought out!
  3. Solid Stuff
  4. A group worth trusting. Never locked you out!

My 2 cents
Andy

1 Like

I must have missed that above… I definitely enjoy Debian and agree with your points.

Sad about CentOS, I believe it has afforded RedHat so much to gather and use in their own enterprise products, but I know they don’t see it that way.

I hope that Rocky and Alma can hold their own for the availability of choice. As it seems to me they are more like Linux Mint is to Ubuntu. A clean up of code and adjustments to make it more intuitive.

But to be honest I don’t know the whole story of Rocky and Alma. Only have looked at them at the desktop level. And have been trying to test out Rocky as a server a couple times.

I do have a possible install that may require one of them due to a specific app that requires a CentOS based NS8 setup, and I am trying to decided which would be the best for that scenario.

I would be open to anyone’s recommendation from those who successfully use Rocky, Alma, or CentOS as their base choice for NS8.

Thanks,

-SF-

If asked, I’d say it wasn’t RedHats desicion. Big Blue called the shots, and they were crying for more blues… :frowning:

1 Like

Rocky is the “official” choice for NS8, even though all are supported…

1 Like