Can I ask something specific for my case?
My home server consists of UNRAID and a VM with NS7.
It serves both as my personal “production” and test bed.
I have installed various modules, but the one module I actually need to be working from day one, is migrate my mail and using webtop (or other web client?).
My mail has few accounts (so I also use LDAP provider and I need to migrate my users), uses smarthost for outgoing mail (which I think is a requirement for NS8 for now) and gets email using a IMAPS connectors.
(everything else to implement in NS8 will be a bonus - I don’t even need to migrate it)
Can I safely move my simple setup to NS8b1? (although I will probably wait until b2 to be honest) Which means a separate VM I guess (and changing my port forwarding etc.).
Again: Move my few users and email setup AND current mailbox contents and re-implement a web mail interface.
In what state will this migration leave my NS7 setup? (asking for security reasons) In other words aside from reverting port forwarding and having lost possible new mail while NS7 was offline, can it be brought back to production?
Future betas and final will need further migrations or actual in-place updates (over NS8b)?
I know probably you guys will not answer this directly, but is there a suggested image (distro) to use? Or to maybe have an answer, are there parameters that can help us choose one image over another? Like more complete or smaller footprint choices? Which is which? Is there any that is more tested?
I don’t think there’s a practical difference among Alma/Rocky/CentOS. I’m pretty sure Oracle Linux 9 can also be used, but they don’t endorse that. I do recall hearing that Debian 11 had an old-ish version of some important package (podman, perhaps?), which would tend to make it less favored. With any of the RHEL variants, a minimal install is all you need.
I am asking about the prebuilt images (that will run on KVM).
Possible differentiating factors:
Smallest storage footprint?
Smallest memory footprint?
Richest-non-NS8 functionality (from start)?
Most NS8 tested distro?
Easiest to admin distro?
Would be nice to know.
Remember that one of NS “selling” points, is to be able to implement specific services without the deepest knowledge of the underlying system (because else, someone would argue, that NS is not needed and everything can be manually configured by an experienced Linux admin).
So such a table would be a great asset.
Ah, then in that case I’m afraid I can’t help–though I’d still be surprised if there were any significant difference (in storage, memory, non-NS8 functionality, or administration) among Alma/Rocky/CentOS, as they’re all pretty much the same distro.
Indeed they are - my question about those three extends beyond NS8.
They seem more like “I want to make my own thing” than make actual difference.
Seeing that NS8 team actually suggests 3 RHEL based distros and a single Debian, I tend to think they actually want us to use an RHEL based solution. So one is “kind of” out already.
Surely you remember the CentOS 8 debacle, where RedHat without warning pissed away the trust they’d bought when they took over the CentOS project, and changed what had for many years been a stable, LTS server distro to a “rolling release” model. Lots of people saw a need for an EL-compatible LTS server distro, whence came both Alma and Rocky (Oracle Linux had already been around for a number of years)–so that’s arguably a reason to avoid CentOS. But among the three, AFAIK, preferences would be more based on philosophy than on any technical difference.
BUT my original 3 questions (and the point of the thread) remain.
(although I now know about #2 probably - seems every service migrated “disables” it from NS7 side)
And the main reason for virtualization is also the fact that you can easily make a full backup - and restore that after migration… (As a saftey measure!)…
It’s not a requirement. You need it only if you want to receive mails from the server from apps and you do not have the mail server installed and configured.
IMAP connector is still not available.
Wise choice: wait the rc for a production server
Except of IMAP connectors, the other things on your server should already work.
It will be attached to cluster until the migration ends.
can it be brought back to production?
Yes.
We do not still know if we are going to have an in-place upgrade.
