Having a bit of trouble with NS 8. I’ve installed it using the qcow image on Proxmox 8. It boots up successfully, and I can access the cluster admin interface.
I get some web socket errors, and the update is still running at 50% in the notifications bar even though software center now says everything is up to date (this persists through a reboot). I did read the other post about websocket errors and NS8 IS grabbing the same IP as another Proxmox container (currently turned off). However, I can’t see how to change the IP in the interface, and can’t access the console. (I’ve changed the IP of the other machine and haven’t had a repeat of issues – though with everything up to date and no migration tool, I’ve not had much to do).
When I try to login at the console, it prompts me for a changed password (as expected) but then rejects the password change with a “Authentication token manipulation error.” The password does not change and I am never logged in.
Finally, my purpose with the install was to try the migration tool. However, I am unable to find it. One NS7 server (an arm-based pi4) shows nothing at all for “migration” in software center. The other (an intel machine) only offers a “firewall migration.” What am I missing? (I’ve tried searching yum including with the --enablerepo=nethserver-testing switch and can’t see the nethserver-ns8-migration package).
On ns7, if doesn’t show up on software center (searching for “migrat” or “ns8”, for instance) you can try cleaning repodata and try again afterwars, or install the package from command line:
The migration utility will only show up, if there are any migratable apps running on that NS7.
An Intel box only running firewall? I’m not really sure anything will show up…
Thanks. I thought that might be the case, but then was surprised not to see it on the intel machine — which is primarily running Nextcloud and webserver both of which I believe are supported for migration.
Hi Mark, the Pi is running Mail and web server for use as a reverse proxy. Both are probably easy enough for manual migration if I can sort the console access issue.
If you google it it pops up a lot of times regarding possible issue re filesystem, shadow file (holding password) and more. So it does not seem to be NS8 specific.
Seems everyone expects the latest and greatest in a pre-chewed VM implementation…
As to NS8, this isn’t actually always the case. The VMs get the least testing…
As a Proxmox user, I wouldn’t touch a pre-chewed NS8, I’d spend the few minutes to spin up a VM the way I want / need it.
As I recall from NS7: The swap partition was automatically sized, depending on installed RAM during the initial installation. Changing this later was possible, but not easy or for the faint of heart. A partitioned Disk is already partitioned, and would need a few operations with gparted to correct!
Thanks for the reply. My read of the docs was that the qcow image was the dev recommended method for Proxmox. Wanting to spend my time configuring my server instead of fixing my own errors, I went what I thought was the recommended route.
Can I suggest that the docs be amended to state clearly that the prebuilt images are only lightly tested and not recommended?
Basically, all testing advancing the use cases of NethServer and the final release of NS8 is very welcome.
Install bugs, migration bugs, whatever itch may be there… Ideally, the team of Devs can smooth almost all out by release day - I also have quite a few clients to migrate…
NS8 over Debian 12 is working much better for me than the qcow image. It seems to be stable and the features work as advertised. So now, I just need to learn the new way of doing everything.
Speaking of, I can’t find it documented anywhere. I know the NS7 mail server was really intended for a single domain and allowed other domains as aliases. But NS8 appears to be intended for multi-domain scenarios. I’ve not gotten to the point of testing send and recieve from multiple domains, though, so I’m not sure.
