My review of some providers I've used for the past years (BuyVM/RamNode/LiteServer/Hetzner/OVH)
Hi, this is my review/experience with random hosting providers that I've used over the past years (BuyVM/BuyShared/RamNode/LiteServer/Hetzner/OVH). I'll only be including ones that I have used for longer periods (at least a few years) so you can give a complete picture of what you might find when you settle down. And because "bad" experiences tend to stick more, I'll be focussing more on the roadblocks than on the good sides, but I'll try to also name some of the good things that I remember.
Disclaimer: This is my personal experience, and some of the things are from my memory. When I move provider I clear the history from my uptime monitors so I can see how the uptime at that specific provider is. If anything is incorrect/wrong, let me know and I'll correct it.
In no particular order;
My first service with them (shared hosting - 1GB) dates back to May 2017. Since then I've had a few services with them. All of them in their Luxembourg/Lux location. So my experience will be with this location only.
BuyShared cPanel hosting, BuyShared reselling hosting (DA), BuyVM slice (512MB/1024MB/8GB), 2 slabs.
My support tickets are pretty "specific" most of the time, the last few tickets were regarding their Path.net DDoS protection and how it would work in a specific set-up. So they are put under "Expert attention" - So Fran has to do them. And as far as I know, he's very busy. These tickets are answered within a few days. But that is to be expected because they are very specific. For general questions (e.g. if I may do portscans on other services that I rent, or unblocking port 25 for sending mail) the response time has been amazing. A few hours for the port blocking, and 2 minutes for the portscan question. Also, when I send a DM Fran almost instantly replies (haven't done that too much tho). Besides that, they have a Discord/Matrix/IRC server that is full of people willing to help (given that you have tried to find out things yourself, nobody likes someone that expects others to do the work for them). So the support is pretty great overall.
This has been my "main" issue with the Lux location so far. From the start (my first services); I had a few shared hosting packages from BuyShared. And sometimes (not often, but it would cause >20 min outages sometimes if I don't remember incorrectly) there was an issue with a service (e.g. httpd/mysql/etc) on the shared hosting servers. This causes sites to throw 500 errors, or just not load. I have mixed feelings about this, because buyshared.net promises; "99.99% Network and Service Uptime". And due to these issues, that is not reached. But on the other side, we're talking about 8$/year packages (with a dedicated IP.. most providers cannot even offer that alone for 8$ a year). So I don't expect it to be great. Nonetheless, it was my experience. Of course, it was up most of the time, but enough for me to not trust it for production systems (who would for 8$ ).
Then, fast forward some time, I have been using BuyShared mainly and needed a VPS for a production service. It was tricky because I did not have any failover for this. It was a DirectAdmin server. I have been using it from 13/01/2020 until 01/02/2021. So about one year. This was an AMD Ryzen KVM 8GB (30$/m) VPS. The experience overall was pretty good, sometimes there was a small outage due to networking issues or something like it. Nothing major, but it is still pretty annoying. I also had some shared hosting from BuyShared running at the same time. And the issue that I've been seeing there seems to be fully unrelated to BuyVM in most cases. My BuyVM server was up pretty much always. But I eventually moved it away because the small downtimes every now and then were getting annoying.
Then we are at the present day, I have two services at BuyVM left; One is a HaProxy VM for one of my services with a DDoS protected IP. And one is a backup VM that has two slabs attached to it. First the one with the slabs; There was an issue a few weeks ago where the datacenter connected too many servers to one powerstrip, causing them to use too much power and shut down. This caused some issues on my backup VM because its storage was no longer readable/writable. Now because this VM is mainly used for backup storage I don't really mind. So I turned off the VM. And turned it back on when it worked again. But (if I'm not remembering incorrectly) it happened again shortly after. Overall this was about 8 hours of the server being unusable for its purpose. I turned it off again when it happened, and the next morning when I woke up, I noticed that Fran had given them a reboot (the VM's) to make sure that the storage attached back correctly. Now, this is an isolated incident, it was not their whole Lux location, just a few slab servers that happen to be connected to this powerstrip. But it did cause some work from my side (because I got Zabbix mails and had to check it) and a few hours of downtime. For me, it is just a backup server, but if I was using slabs for production stuff (a CDN or something) then I'd be pretty stressed about it.
And for my second service; My HaProxy load balancer, I have a Path.net protected IP. Overall it has been great. I do get a Hetrixtools notification every now and then (2 times a month or so for <2 min downtimes - Timeout). Also, a few days back, the subnet that my protected IP is on had an issue on Path's side where it would stop announcing due to an edge-case, causing 40 mins of downtime for the IP.
Besides those issues, the Lux location has been out of bandwidth for a while, so at some peak times it will congest, and it causes packet loss. They are working on it (still waiting on GTT) so I hope that will get better soon. I'm not sure how much impact it has on my VM's as Hetrixtools monitors from many locations, so I don't think it get's triggered with there is just packetloss.
Frantech/BuyVM/BuyShared (RIP - No longer sold) has been great, the provider is not expensive, the support is great, the Discord server is fun to read sometimes, and Fran seems to really care. I've seen him help people for long periods of time. And he won't kick you off for lame reasons - Very pro-privacy. They had an issue with carpet bombings a few months back, and Cloudflare wasn't able to help them, and instead of listening to the demands of the attackers, or paying up, they got in contact with a DDoS protection provider (Path) and got it sorted. They do everything they can to keep their services up and running and help their clients. For my taste, there is just a little bit too much risk to put production stuff at BuyVM when you do not have any automatic failover to another location as I've seen longer downtimes happen. I know 100% uptime is pretty much impossible, but the issues I've seen are too long to keep anything that must stay up with BuyVM. It is great for off-site things like backups, or redundant services.
My first service with RamNode was a 128MB VPS, that I used as a VPN. It was invoiced in July 2015 ( I still renew it to this day ) - Just not as a VPN, because it's way too small, but as a "Bastion" server to connect to my other servers. All of my services have been in their NL location.
VPS (OpenVZ & Openstack), cPanel reselling hosting.
The support I've gotten at RamNode has been great almost every time - I cannot remember any bad experiences but the human mind isn't perfect. Most questions are answered within the hour, quicker in most cases. Their support is just top-notch. A few weeks ago I had an issue, I tried extending my partition (LVM) and I removed the SWAP partition and removed it from fstab. But afterward, the VM failed to boot. This is a VM that cannot easily have any off-site redundancy because they were client websites. So after it failed to boot, and the time reached 4am, I decided to restore a snapshot I made before starting. But it failed to recover every time. So I contact them, and within minutes I got a reply. After a little back/forth the restoration process was doing its thing. But it failed again (it took >30 mins every time). So support started the restoration process and I asked if they could monitor it. So they set a timer to check if the process was going. They checked if the server started up properly after the restore, and if the DirectAdmin panel was listening. I went to sleep and set an alarm for a few hours later (7 am) I woke up, opened my phone, and their latest reply to my ticket with a message that everything is working correctly and that everything is UP popped up on my screen. I turned it off and went to sleep. Just perfect support, and it hasn't declined over the years, it is still amazing.
The uptime has been really good, sometimes services go years without any monitoring software reporting any downtime. And when downtime happens (a few weeks ago there was something wrong that causes some small network issues across all of my servers with them) it was resolved and caused <20 mins of downtime. The only notable downtime that I remember for the network was when they were attacked a few years back. During this time, they put everyone behind their remote DDoS protection provider to protect the network. But it did cause some issues.
I did have one "faulty" server or something like it. I ordered a new VPS to use as an "offsite" server for things like monitoring. And it had some issues. On the second day, it went down for 10 minutes (the host rebooted). And a day later I had some issues again where I could not ping it. Because I did not set-up anything I just requested a refund in credit so I can pay my other services, and moved that VM to another provider. It was easy to do because I did not yet set up anything on it. But from my experience, this is not the norm at all and seemed to be just that host node.
RamNode is my go-to provider for reliable servers, and I'd trust them with hosting production systems. They do not have live-failover from what I've seen, so you are still subject to the uptime of a specific host node, so if the motherboard of a server would die, then you'd be down. So I'm experimenting with providers that use SAN storage and redundant computing nodes. But if that is out of budget, then RamNode is perfect.
My first invoice with LiteServer was at 10/05/2017 for a small 128MB VPS. I'm currently using them mainly as a backup provider for some services that I have. For example, I have a bastion SSH proxy server at RamNode, and I have a backup at Liteserver in case RamNode would fail (because I would not be able to access anything without connecting to my OpenVPN server at OVH/Hetzner and manually finding the IP's).
VPS (mainly 128MB servers, and one storage VM for 3,75 EUR/m)
The support has been good, I have not created many tickets, but they'd always reply quickly. I even created a ticket about a service that was sold out, waited a few hours, but then bought it at another provider because: It was cheaper, and I like it cheaper. And because it offered more for the same pricing (talking about BuyVM slabs here). So I closed the ticket. But they still replied to it (saying that they know I closed it) and that if I still needed it, they'd be able to provision it manually because I already had some services with them. So overall the support has been good, I cannot give any more specifics because I did not create many tickets.
The network has been great in the years that I've been with them, there is rarely any downtime. They recently upgraded lots of their infrastructure, so it should be even better. I can only remember one "major" outage that lasted for, I think, an hour or so due to something being broken. This happened some time ago, and I think can no longer happen after their network changes. I also noticed some packetloss during mid-April. Freshping notified me a few times a day (I have it set to notify when even one request fails). And I could see the same packetloss from my connection. But this is no longer the case, and lasted for a few days. I'd trust them to host some production things without failover.
I have not used them for anything important, just for redundancy if another provider that I use would fail. But they have been great. Support is responsive, the network is good, and the servers are priced pretty low and work as expected.
- Body was too long to post so read the rest in my reply.