Honestly I totally disagree, this community is about avoiding SaaS and controlling the important part of the stack to keep your own data under control.
If the service you need is just a provider of vm for your own business, is not a contradiction and it is still selfhosted as long as you control it.
Check it again, I used it for years before switching to something with more features
If you have a prosumer router I suggest you to use the ddns in the router plus a reverse proxy. This would be the cleanest solution.
If you can not, once everything is working with your external access to the synology, the dsm has a built-in reverse proxy so it can redirect http requests to another server. Although this proxy is really simple and limited it can get the work done if you setup is simple enough.
Got it.
But looking that the question has been posted in a selfhosted thread I don’t see the recomendation of the cloud solution :)
Why?
There are not benefits in the cloud solution over the hosted server. It could even be more expensive on the long run…
And yes, kasm has a cloud solution in its web page
Tailscale too if you use the fee implementation Headscale
And in both cases you need a vps fully reachable from anywhere
Check Mandos , if you are able to secure enough the server (inside a safe box?) then you are good
Explore kasm, it is fucking amazing Kasm
Performance wise AMD and Intel run circles around the m4, there is not any compititiom here.
Performance Watt per instruction is where the m4 really shines and still I have my reservations
Plenty of ipv4 only in 2025?? Really?? Without a possibility to activate dual stack or just the dslite-crap?
I honestly didn’t think this could be a real issue in those years.
Small and stupid question.
Why don’t use a ddns client to update your ipv6 evytime it changes? With a ttl of a minute your shouldn’t be able to see any downtime…
I (genuinely) thinks you are trying to solve a small problem in the complicated and hard way… M
The problem with such advance Sw is the overwhelming list of options and the lacks of sane defaults
It is not the same to find 10 different (and complex) solutions when you are evaluating what you can do for solving a problem. It adds more noise to the solution than anything else. And of course the minimum resources needed ;)
For the downloading I suggest you to have the download folder and the main storage both exported under the same nfs folder. Quite handy.
It depends of what you want as future proof (expansion capabilities). Usually home user nases come with low power cpu, a high power cpu usually is a enterprise grade nas, really costly for a home user. So having it separated makes the cpu upgrade easy but now you have 2 boxes. But if your terramaster comes with a decent cpu I don’t see any problem.
True nas scale is really a behemoth able of almost everything. I would start with something more reduced like omv or unraid. You really don’t need the advance enterprise features of that and it will add only complexity to the setup.
If you use NFS for exporting folders from your nas, the “computing box” will see this as a local folder, so no need to have 2 copies of the same file.
Hope it helps
Huh???
Honestly I don’t see your problem, a nuance? Sure! An unsolvable problem? For sure not.
If you want to have your system reachable from the Wan then you will need a domain name. If you have a domain name then it is needed to be resolved by a dns server.
If there is a dns resolver then you would able to update it dynamically every time your ip changes.
True that the time alive of the dns records must be low enough to ensure that an ip change does not let your system down for an unacceptable amount of time.
You got it quite well :)
Have fun with the build
Yeah, but even if they seems to be contradictory messages, they are not.
RAM is always the first resource that is depleted is a nas/homeserver built, more ram or the possibility of expanding the ram is always a safe bet. But for a nas or a non Realtime system ddr4 is also not mandatory.
Reduce the cooling goes into the direction that the system is not going to be under high cpu load, so no need to dissipate, you can even reduce the cpu.
SSD is not necessary for a nas but it will make you VMs or containers snappier.
So coming back to the subject, your cpu is overkill, and therefore your mother board too (it has even support for sli!), having more ram is always good but my opinion is with 16gb you are good to go for a good amount of time (my server runs on 8gb and I don’t experience any problem yet), obviously, your cooling needs to pair your cpu tdp.
SSD or m2 memories are cheap additions, and like somebody else suggested, it is good to have many pci slots even if they don’t go full speed (x16) but will give you the flexibility to add more sata ports or Ethernet connections on the 10gb.
Honestly, if budget is not a concern you can not be mistaken with this built, but I can see potential to spare like 100 bucks in hw and another 100 in bills in next years
Honestly, I have the impression your setup is oversized (knowing nothing about what you want to run)
NAS systems set on idle like 90% of the time unless your are doing really crazy things with de duplication and distributed iscsi for super big volumes, that I have the impression your are not going to do.
You can probably cut the performance/specs to the half and still being good for the following 10 years and the extra that you will save on electricity too.
As a comparison I checked this built against your synology and in the multicore setup is x10 more powerful (https://www.cpu-monkey.com/en/compare_cpu-intel_celeron_j4025-vs-amd_ryzen_5_5600x)
Just my two cents
You will need to explain a bit further this statement to mild knowledged internet stranger…
Because the point of waf is exactly about reducing the exposed surface…
A nas is your friend.
And there is plenty is space in the Caribbean Sea
I’d Lineageos solving the problem of the camera’s quality and the banking application refusing to work?