backup option isn’t necessary. you can configure backups in Discourse admin.
just follow the basic standard install and don’t overthink it.
With or without those apps, I think you’d be able to install Discourse.
Is this a concern? WWYD?
Heard some issues about some of these providers using very old CPU models like from 13 yrs ago. There’s obviously no assurance about what you get in this context of vps, as they depend on many factors including vps location, & providers are rarely even open about available cpu models in these contexts
OVHcloud support itself said their min plan (that i’m to go for) with the factors I specified, would likely result in an Intel E5v3 (2014)
Contabo doesn’t wanna say anything abt it. So this can only be speculated using other sources, pretty much all of which suggest they’re much newer like EPYC 7282 (2019) (as per LTE) or even newer ones
For those needing, this source tries to show what CPUs may be used (can’t be too relied on) Contabo CPUs | VPSBenchmarks
It is a virtual private server. It is a ‘virtualized’ environment. The actual hardware runs numerous instances. It is absolutely sufficient for running the operating system they are providing it with along with whatever you choose to install.
That’s what people do with these things, they run web servers and mail servers and database servers and RTMP servers and who knows what. Obviously at some point in time you will run out of some resource, whether that is storage space because people are uploading huge video files or RAM because you are running some intense processes or bandwidth because millions of people are hitting your site every minute. At that point in time you will have bigger problems to worry about then the cost of your vps. Most of them are up-gradable.
You should be able to have your forum up and running less then an hour after you pull the trigger on the vps. Register your domain, get your DNS records set up and install Discourse per the standard install and post a link so we can all admire your handy work
Contabo VPS provisioned Subdomain pointed to IP & propagated Email DNS propagated & smtp credentials on hand
My local pc is Windows so I need PuTTY right? After starting putty, the following pops
1) Do I Accept/Connect Once?
2) Here
is it meant before/after install? At which point exactly?
3) Do I need Filezilla before installing Discourse?
4) (I’m aware this is the 1st step) On my VPS control panel, I can either select only Ubuntu 26.04, or 24.04 with 1 free app from the followng. In addition to the apps in pic, the dropdown lists •Paperclip •Bitcoin Full Node •IPFS Node •Webmin •Proxmox VE 7 •Horizen Node •Ethereum 2.0 - eth-docker image •Ollama - phi4 •Flux Node
ask.discourse.org’s AI said most useful can be Dokploy & OpenClaw & lotta others are irrelevant/better avoided (response below). Consider their impact on server resources & whether they can be free. Considering all, is even 1 worth includin or nothing’s better?
Nah, use the free and modern Windows Terminal app from the Windows App Store.
I mean, since it’s optional, I would say just skip it. These are not needed to install or run Discourse.
You’d do this after installation. Just do the install first. I see you’re brooding a lot on what to do pre-install, while that’s okay, I would suggest to just go ahead and do it. It’s not as scary as it makes out.
you can use either as long as they are LTS. i have a site running on 26.04 and another on 24.04. in your case, i’d suggest the latter one since Discourse is known to run on it without any issues at all.
no, use terminal, just like Robert said.
what does the standard install doc say?
what does the standard install doc say?
please follow it and stop adding more complexity that you do not need. if the standard install doesn’t mention it and you don’t know what it is, you don’t need it.
I’ll add that while the community is always keen to help, if you feel you hit a wall regarding the required knowledge and technical skills and you have a budget, you can also post in Marketplace for tailored help 
Almost all of those things will probably break things. If you’re an expert at whatever panel it is and Discourse, then you can make them work together with a few hours work. And it’ll be likely to break Discourse again when the panel upgrades itself. You really don’t want any of those unless you have a very specific need, and you don’t.
All smooth & at the end of the terminal’s questions asking for credentials. What should I provide here? This was a Q I had earlier
It has auto completed with an email domain based on the subdomain (noreply@forum.mydomain.com) that I don’t think I really have. I have a few aliases though.
1) Should I provide a real alias, or isn’t that pseudo email address better cuz it’ll be truly no reply or smthn?
2) Also, the connection to the server doesn’t last long & gets time out. SSH ing again from the next line doesn’t connect, tried few times. Is this normal? Is it ok if I close the terminal & SSH? Will this have ruined the installation or can I continue from where I left?
there’s usually a redirect email section in your DNS and you can add a catch all and choose which email address to forward to …
I don’t think that’s normal - mine doesn’t time out - although you can configure things to do that but unlikely the default - are you running out of memory or disk space by any chance? remind us what the spec is
Most likely you might have not configured SWAP and you are running out of memory …
Here’s a guide to do that:
My Hostinger support said it’s normal to timeout after some inactivity. Tried after a few mins and, the Discourse setup pretty much resumed. Then following were shown. Seems subdomain’s connected
Then after many lines showing discourse being downloaded, here’s a part of the last 2 command lines which stayed as is for a while (later connection timed out so dunno if other lines followed)
1) Was the build part on terminal successful?
When I checked my subdomain link on browser, it showed a browser error page (‘This site can’t be reached, forum.mydomain.com refused to connect.’ with ERR_CONNECTION_REFUSED)
Hostinger sup said this is v likely cuz DNS ain fully propagated yet (I pointed it ~24h ago)
My troubleshoot with ask.discourse.org's ai suggests negative news
When I ran those brief commands from the ai they both said -bash: ./launcher: No such file or directory
Building is not inactivity!
Consider using tmux with that provider.
Plenty of good advice here. I followed the standard and got in a muddle with a few aspects so I asked claude.ai to help me. The outcome was entirely satisfactory.
Troubleshooting this with Contabo is just unacceptable
Contabo’s live chat ain live so the current feedback loop using tickets is unreasonably delayed & inefficient. What should be a chat/call where many things can be solved in minutes, this way; with hours between literally each message is so unreasonable & like using pigeons, and solving each smallest thing that should just take minutes ends up taking DAYS. This is compounded by each of their response not being that supportive either
[My vps specs: 4 vCPU Cores, 8 GB RAM, 200 Mbit/s Port]
The disconnecting issue in brief
The moment it disconnects nothing’s shown on terminal, so don’t know in how much time it happens. Only on input do I find out; on input on the command line root@xyz:\~# , it doesn’t show any input & immediately populates that line showing this;
root@xyz:\~# client_loop: send disconnect: Connection reset
PS C:\\Users\\user>
I’ve checked after 5 minutes of inactivity, & it has disconnected. I’ve checked after 3 mins, it had been connected. Based on which, it disconnects sometime bw 3-5 mins. But I’m not sure if this is the same when an install is running.
When I asked about this from my Contabo
[hours..]
without much explaining they just migrated it to another host system ‘to improve the performance of the VPS’. I tried again on the new migrated one & that too had the same disconnecting issue! Sent a ticket again
[hours]
when they responded they answered a lil more, agreeing this issue is unexpected behaviour & that when they pinged my server it shows quick responses without any package loss & that the issue maybe due to a network config/issue. They’re troubleshooting this now.
Given the above, (the ram available, their diagnoses for this unexpected behavior & current troubleshooting) I wonder if I should do any changes yet like adding swap files? (kinda understand it but also a lil complex for me) maybe I’d first check if they set this back to normal working order
8GB is a lot so maybe not the issue.
You can always inspect memory use via vmstat or install something friendly like htop.
This sounds more like a network issue tbh.
Great that that’s in line with you too. I’m chatting with them now