MG2R

Well I’ll be damned. Never heard of that. I used to run LVM on top of mdadm back in the day. I’ll have to play around with that.

While I haven’t used it myself yet, my understanding is Ceph could offer you what you want. It does come with a whole host of other issues though, like needing a relatively large amount of compute resources and it not having raid5-like efficiency. But you can add drives at any point in time and have it auto rebalance online. Edit: if you’re ok with not going drive by drive you can go for either LVM on top of multiple arrays or an all-in-one solution with ZFS, where you can add new arrays to the pool. Of course with both of these solutions you do not get automatic rebalancing meaning that your performance will not necessarily increase when you add more droves.

Everything was set up correctly. It’s just a symptom of shared IP addresses with mail servers. Not Mailgun’s fault really. They had anti-spam detection from what I remember but nothing is 100% watertight. as soon as you pay for a dedicated ip, those issues go away. Again, to their credit, any time we noticed such an issue they had us switched to a different IP immediately.

Used them at a previous employer. At that point they were our cheapest option and it showed. We had it happen multiple times that the sending IP we were assigned was shared with someone who sent spam and thus our mails were getting blocked. Of course, if we would’ve paid for a private IP we wouldn’t have had those problems. To their credit, they were responsive on their support and we were never annoyed enough to switch providers.

One big detractor for Unraid for me is the fact they use a proprietary filesystem. Meaning, if there is a catastrophic problem with Unraid at some point I can’t just plop in a Linux drive and work on the filesystem using readily available tooling. how likely that may become a problem, I don’t know. It’s just something to consider

Well it runs Docker in the background meaning you should be able to run these things in other OS’es as well. a lot of those in the list are by a user called ich777. You can find their work on the upstream Docker Hub as well: https://hub.docker.com/search?q=Ich777 now, it’s quite possible that the experience is more polished, one click and go, on Unraid because someone put in the work to figure out the right parameters for the container runtime for you. Not sure if that’s the case here. but in general, TrueNAS Scale and Unraid are both simply Linux servers with a flashy Web UI so there should not be much differences in capability. Question probably boils down to how polished of an experience you want to have vs how much tinkering you like doing. As I said: I’m of the tinkering variety and I always try to understand the underlying technology and components of a tool I’m using. So I tend to ditch the pre-packaged fancy web UIs and just run Linux + Docker/K8S myself using the command line. I understand not everyone is this way Cool thing is that TrueNAS scale and Linux are free, and Unraid has a thirty day trial. So install all of them and see which you prefer management wise.

Let’s try @alpenwasser xD

Not quite sure. I have not run unraid myself so would not be able to tell you it’s strong points. I tend to go the diy option for most things so I just run straight up Linux. TrueNAS scale is basically pre-packaged Linux + Kubernetes + ZFS with a fancy ui on top. can you share where or why you got the understanding unraid is better in some way for hosting game servers?

Like the multitude of other threads: run TrueNAS for a point-and-click, easy-to-use experience. Install your favourite flavour of Linux then run either Docker Compose or K3S for your containerised software if you want the CLI experience. Whatever route you take, make backups first, dink around later. I like Borg Backup but there’s so many other ways to do it. Seriously, back up your stuff somewhere _before_ playing around with your NAS.

With that selection of drives I’d run my OS off of one of the SSDs, then set up a zpool consisting of either a mirror with those 500 gig drives or a RAIDZ1 with the 4 drives. Then add the second ssd as l2arc

OP, keep in mind the bathtub curve. Hardware will usually either fail when it is almost new or when it is fairly old, with little failures in between. https://en.m.wikipedia.org/wiki/Bathtub_curve To me, it sounds like you had a bad batch of drives. And it sounds like while working in your system you disturbed a drive that now has performed reallocation. @LIGISTX is right here. Burning in a drive is definitely good practice, especially if you buy multiple drives at once. If they’re from the same batch and there was a manufacturing issue with that batch, you are jeopardising your data by putting it on a pool with a high chance of simultaneous failure. About cascading drive failures: this is anecdotal but one of my first gaming rigs back in the day just ate hard drives. We’re talking every few months, it’d have a drive die. Back then I never figured out why and I ended up getting rid of the system. my running theory now is bad power. Others have commented on your PSU but the sticker tells you only how many Watts it’ll be able to deliver. It tells you nothing about the quality of that power. Too much ripple in the voltage can be devastating for sensitive electronics like your hard drives. The unfortunate part is that there’s no easy way to test this without an oscilloscope.

Ok, so interesting. OpenZFS (what TrueNAS uses) is different in this regard than oracle ZFS. so it indeed is not as flexible as I remembered. I think I had been reading the oracle ZFS manual, which does not stipulate this problem.

Again, yes there is. You can remove a vdev and add a different one. Yeah, that’s my point. I just wanted to make the point here that it most definitely is an option and that it’s not hard-set that you MUST start out with the pool exactly in its final form and CANNOT upgrade or change it later, which is the message I got from your posts. that’s my point in its entirety: it’s an option, even if it is not the prettiest one for this use case

Congrats, you’ve run into someone with a similar interest but a different viewpoint on said interest. I’m not looking for your posts, merely finding them among the threads I’m interested in. This is great. We can both learn. Everyone thinking alike gets us to a pretty boring dystopia. Vdev yes, zpool no. You can have multiple vdevs in a pool and add or remove them as needed. Not all vdevs need to be identical nor do they need to be added from the get-go I started with 2X2 TB mirror, added another 2X2 TB mirror, then added a 3x4 RAIDZ1 as well as a hot spare. Then replaced 2 of those 2TB with 4 TB drives as one of them died. That currently leaves me at 14TB usable where I started with 2. Is it ideal? No. Was it seamless? Heck no. Is it an option? Yes.

True. My point is that it’s often not needed. Once you’re standardised on Nextcloud and have your camera roll syncing, what is there still to back up? My phone settings I can perfectly replicate in about 15 minutes if I ever need to start fresh. My messages, WhatsApp or otherwise I specifically turn the backup off for. I really can live without them. Even if you want to enable backup for those services with iCloud, the grunt of your data will be the photo roll which can be handled by Nextcloud perfectly, meaning you won’t need the paid iCloud services

If it MUST be on-prem, buy yourself a Synology box and then buy yourself some backup solution. In your case I’d go for managed cloud storage. Your core business is not running servers. Why waste your valuable time doing it? Pay someone to do it twice as fast, twice as good at half the cost. oh btw. Get backups. Then backup your backups. Sprinkle some more backups on top. Backups!

OP is asking to get away of cloud server reliance and you’re telling them to buy a cloud service. Amazing. Nextcloud integrates nicely with iOS and can automatically capture images from the camera roll to sync to your cloud drive. Calendar and contacts are all perfectly integrated as well. Been living cloud free like this for years now. @OP you’re saying you want to replicate services on each other’s system. That requires the other party having access to your data. A service cannot serve data if it doesn’t have access to it. If the other person has super user privileges on that machine it means they will be able to get to that data too. id forgo multi-site services and simply offer backup space to each other. ZFS replication can work but consider you would need to encrypt that. I personally use borgbackup for my needs. It provides encrypted, snapshot , deduplicated, compressed backups for your data and generally works without a hiccup once set up. I have three years worth of snapshots to guard against accidental deletion and cryptolocker type attacks. software wise, for your cloud needs. as I said: Nextcloud provides you with a mostly turnkey solution to self-host cloud storage, calendars, contacts lists, reminders, notes, etc… they even have their own chatting and voip system these days. you can set up accounts for each member of the household and even use client-side encryption so the admin account does not have access to the data. you can federate Nextcloud instances so you can share directories with members of different instances easily. it can integrate with collabora server to get Google-docs like online office suite. for access: consider a service like tailscale or zerotier. If you don’t like a cloud service for VPN-like secure access, I would set up Wireguard. It is the easiest VPN solution by far. It does not require complex certificate management. It is the first one to by seemed worthy to be added to mainline Linux kernel by Torvalds. Thus, it is ver performant and requires minimal resources. The iOS integration is absolutely seamless.

Disagree. TrueNAS runs ZFS which definitely has upgrade paths. You can’t change a vdev as in the amount of drives, but you can definitely upgrade the drives and get capacity increases that way. I’ve been doing it for years. of course, you only get the cap upgrade if the smallest drive in the vdev is upgraded and until then you’re “wasting” the space of the bigger drives. So there is a reason for Unraid here, but saying that you have to build your entire storage array exactly as you need it for years to come with ZFS is false Reading some articles, it seems this has a very different use case than general purpose storage to back compute. I’d not use it for this case, although I of course might be wrong as o don’t have hands-on experience That’s pretty darn vague. Containers are the most popular technology for running web services in the data center. I can almost guarantee you that all large “cloud drive” offering run on containers people have issues doing all sorts of things, doesn’t mean the tools are to blame. that said, containers do add a layer of gotcha’s to working with Linux if you’re just starting out. However, most complexity depends on your specific management system. Simple Docker is the most popular for a single host. Probably using Docker compose to actually keep your configuration in code. I’d invest time in learning the basics of Kubernetes. Run your container workloads using that. It really is quite powerful. ———————————— OP, you sound really, really new to all of this. It’s great that you’re looking into all of this but please, for the love of anything that is holy, look into proper backups! raid is not a back up and at some point in the future you will have a catastrophic issue and you will lose data. Use proper backups! really, backups are what you need to look into first. Get those sorted before dinking around with new technology for your storage. I like Borg backup, but there are many other solutions. Seriously backups! did I say backup enough? backup.

1. Different hard drive sizes I think the only real solution here is unraid. Otherwise, run ZFS with two 4-drive RAIDZ1 vdevs in a pool. Each vdev will be limited to the size of the smallest drive in that vdev but as you swap out drives the vdevs do grow automagically. 2. Software im running Nextcloud and its database together with Plex and a whole host of other services on Docker Swarm (deprecated clustering system built into Docker). I don’t see what you mean that running NAS software in containers ain’t a good idea. tbh these days I’d look into something like K3S or a different single-node kubernetes solution. Kubernetes is just so much more mature and flexible vs a simple Docker installation. it does take some getting used to though.

My experience is TP-link is not value for money. The failure rate when I was still using their products was incredibly high. Netgear these days does idiotic account requirements where you need a Netgear.com account to manage your device. Otherwise their hardware is mostly solid

Money is absolutely an issue for servers. also not all servers will actually run OS storage.

There is a way. You’ve linked it. why do you think those kits are shady?

There's another layer to this I missed entirely. This "performance analysis" has nothing to do with the hosting directly. It's looking at website render performance in the browser mostly. If you keep the same website but host it yourself your metrics won't magically improve. Keep in mind: for me here Belgium your website loads just fine and is snappy so read from that what you will. If it is dog-slow for you then it's probably because wix does not have a server close to you. Same conclusion as before: try a different hoster, maybe one specifically aimed at Indonesia.

oh wait, I misread this. Didn’t realise it said test server location. I thought it said simply “server location” and figured it was the wix server you were connecting to. Sorry for that. Regardless, what I said isn’t untrue. Say you were to host your server in Australia, you would still have to contest with physics for clients in Bali: according to https://wondernetwork.com/pings/ simply setting up a TLS connection is already three round trips before you’ve even started transferring the first bit of actual request or payload data. Now, I’m not saying Wix is a superb host or whatever. The only thing I hope to really get across is that there’s a reason most businesses don’t host their websites on a self built computer in a closet under the stairs. In your case I’d look for a different hoster if you’re unsatisfied with your current. if you want to go for it, by all means do. I mean the traffic you’re talking about will probably be able to be handled by a raspberry pi. Just know that you WILL have downtime at the absolute worst possible time and you will hate every second of it at that moment. Otherwise self-hosting is a fun learning experience. I’ve been self-hosting as much as possible for years now. Even hosted my own mail server for years. They put me in the mad house for masochistic tendencies when they found out.

This is such a bad take it’s hard to even begin breaking it down into something reasonable. All services will work with law enforcement, eventually. If they don’t, they will be shut down, eventually. The question becomes what data does a provider keep track of by default and what can they be compelled to keep track of legally? in the case you linked, they received a legally-binding request for data they were legally required to obtain. Kinda hard to say no to that now is it? Saying “they did this one thing which _may_ be _misconstrued_ as being against their privacy claims so you might as well switch to a provider in the states with the worst privacy record on the planet” is idiocy at best and insidious at worst. You are actively try to promote a worse privacy solution for everyone because the needs of a single person were not met by a certain service.