when it clearly should not be doing it.

No, sorry. We don’t know enough to say this. What were the options when configuring ext4? What record size is grafana using? Mariadb or postgres? Is there contention with another db? Is there a mounted NFS path in the same ext4 partition with weird block settings? There are lots of legitimate config reasons for op’s problem, but there are so many parts of op’s setup we don’t know, we similarly can’t possibly know what the issue is.

Swappiness works as intended in most default scheduler configurations and seeing high swap is a symptom, not an issue. It is almost always more productive to find out why the swapping is happening.

Not sure what you’re getting at with TRIM.

I agree that hitting swap is not ideal, but messing with swappiness sysctl is almost never the answer and very often leads to more problems than it solves.

If you want to tune for workload without needing fairly deep understanding of how Linux manages memory, use a sysctl that sets a whole system behaviour like CPU governor presets.

IMO, Linux is configured for “old” systems by default: slow HDDs, and constrained RAM pools, where disk IO really, really needs caching, and where idle background processes take a large fraction of RAM.

Not really a matter of opinion, and not true since kernel 5.1 when ssds became first class citizens.

Ooof, that domain name, tho…

So, couple things here:

1. Most VPS have elastic memory, meaning they can borrow burst ram from the host for short periods. This means you 1GB vps can probably run at 1.5GB or even two during some demanding jobs for a short time. And depends on the provider.
2. You don’t document how you installed any of these.
3. .net core doesn’t work that way, it’s not “running” underneath the arr stack. There is some init stuff that happens when you first start a .net app, setting up DLLs, paths in /opt and such, but there isn’t a resident “layer” involved.
4. My arr stack executables never, ever exceed 700MB, 2GB is more than enough and pretty conservative. I’ve watched these metrics since the days when radarr and sonarr were backed by mono. You do need some mem space to store temporary show/movie metadata, but it’s trivial.
5. As others have said, peak demand is what you want to know, not base requirement.

Generally, I like where you’re taking this, but I think some more experience with how processes and memory work will help you a lot.

Opnsense is based on FreeBSD, not openbsd. They are different in handling jails, driver and fs support, so it does matter if you’re looking for info.

I appreciate your candor, but do you understand why the backlash happens?

One thing I wish ai would teach vibe coders is how to use semantic versioning.

Trying to infer state of project from versions in ai-assisted software is an exercise in confusion.

Agree on the Jonsbo. I have an N1 and it is not great for cooling HDD either. Jonsbo seem to design their cases as ideas and never actually test them.

I have found that a combo of Terraform and ansible cover almost everything, if that’s of any use.

Will test, thx. Keep up the good work.

I initially set it up because both sonarr and radarr were making their own api calls to indexers and I would get some “daily max exceeded” for certain sites. The idea was that if one host made the calls instead of 2, I wouldn’t get rate limited as much.

Now I only have my nzb and pb indexers, the first has no such limits because I pay for it, and the second is only used as a second priority, so I don’t really hit limits anymore.

If you have a bunch of arr services, this is useful. But I only have the two, so I don’t think it’s much work to set indexers up individually.

Edit: this ^^ is my suggestion. Add the indexers manually.

To me, it’s a good stack. A lot of ppl have lower numbers than this, but then they set up a VM with docker and just run 12 extra things there. :P

Yeah. I have a dell r430 sitting unused and a supermicro xeon-d board, both have a BMC.

It’s really nice to have, but server stuff is expensive and power hungry.

Plus, I’ve shifted my way of looking at the setup where I’d rather it just default to starting back up properly. Even if I have a failed disk, if I’m away there’s nothing incan do about it until I’m physically back.

I’m almost 30yrs in it/devops/infosec and I’m no longer interested in owning an excavator just because I can.

Incus, all services running in lxc. Ansible and templates to run startup stuff for new containers (ssh keys, etc). My docker setup is primitive, I regressed down to individual containers for each with docker compose. It’s so trivial to restore last night’s lxc backup that k3s/k8s doesn’t really help me. In fact, i’ve been simplifying everything to suit my use case, including hardware.

• sonarr
• radarr
• bazarr
• prowlarr (I’m close to getting rid of this)
• transmission
• sabnzbd
• kavita for comics only
• calibre Web automated (CWA)
• frigate NVR
• navidrome
• immich
• openwisp to manage 2x openwrt access points
• homeassistant
• technitium DNS x2
• uptimekuma monitoring x2
• ntfy for push notifications
• ansible
• jellyfin
• radicale
• syncthing
• Wireguard
• wanderer for GPS track mgmt on “big screen”. I track my GPS trails and drives as a hobby. I’d love a self-hosted comaps web client, but I can’t really find one. :(

Plus lots of transient trials that come and go with interest. Rotating 7 day/4 week/3 months backups by NFS with pruning, includes snapshots, but I set expires on these.

This is running on 2x minipcs with 48GB ram each, but they are way overkill. I just received a mini itx i5 board, which will house all of the containers, save VPN and 1x DNS, which I will run on an SBC.

So I’m going down from 2x servers + 1 NFS nas to one 1u server for everything plus one SBC so I can reboot it.

Power consumption is king factor for me now.

I suppose so, but all they’ll see is a an encrypted session.

I haven’t found one, no. I just use markdown and a static HTML site generator.

Is this a trick question?

Websites are complicated. Easy, secure, cheap: pick two. There is no such thing as a universally easy, secure and cheap solution, or else we would always be using it.

Assuming you want secure and easy, I would suggest you pay someone who knows what they’re doing.

It was time in 2013.

It isn’t technically tough to do anymore, but you can only self-hosting the sip endpoint, you need to rely on and pay for upstream trunking.

Also, be aware that 911 or emergency service contact is all over the place. In Canada, VoIP providers do not provide 911 unless they have a pots tie-in somewhere, and there are still tragic issues where 911 calls are routed to dispatchers thousands of km away.

If you have any nontechnical people who expect all the features of a landline using this, don’t do it yourself.

You don’t even need to lift a finger. I use WG Tunnel, which turns on my VPN as soon as I leave my WiFi. I’m always tunneled or home.