New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
@jimaek mentioned somewhere that he's running them with 512mb ram or more.
Right now im running 32 probes and many of them with less then 512mb ram. Smallest one is from @Cam with 200mb of ram and its also working. Only docker, globalping container and a monitoring agent.
https://globalping.io/users/gbzret4d
Damn, you are on fire! I did not know you can register, will do mine one day.
Just use my adoption token 😁
how does one get alpine installed? I am using the only iso mount opiton netboot.xyz, but it fails due to getitng the error "no space left on device"? when setting install options like passwords, and network config.
This correct [↑]
Be sure to buy.
Here's one from my yesterday's session (also NAT VPS):
Uptime could have been much longer, if not for my stupidity.
I couldn't resist and ordered 3 smallest packages, 1 in each location. Could use for a personal VPN and for some backup transfers (I have scripts that use rclone to copy backups from one cloud storage to another for redundancy).
@Cam, I think you're open to some feedback since this is a new CP?
I didn't have any prior experience with Gullo and I'm not familiar with the environment used with your other CPs. So the notes below are based solely on the first impression from the offers in this post with the VirtFusion CP (which I've used with other providers).
While what I'm going to mention next might sound like complaints, it's not. I'm absolutely not complaining. Can't complain at all at this price point.
So far, the VMs have been running smoothly, with good network and I/O performance. They're worth more than what they cost on this offer for sure. Please treat the notes below more like suggestions or a "wishlist" of things that might be improved in my opinion. Other than that, I'm really happy with these boxes.
First of all, I didn't expect the Debian templates to come with XFS as the root filesystem. Almost everywhere else, the default is ext4. I personally try to avoid using XFS for my own reasons. I'm fairly experienced with Linux and know how to work around this, but many customers might expect ext4, as almost every guide or script for VPS tasks assumes ext4. This shouldn't make a difference in most use cases, but in rare cases, it could cause some issues. I'm not trying to start a holy war of personal preferences here — I just mean that ext4 is much more common, and it might make sense to offer it as an option.
Another thing I noticed is that once installed, Debian 12 from this template doesn't have any locales built. If you run an
apt installoperation, it might complain about this and suggest runninglocale-gen. Butlocale-genwill fail on a 384 MB VPS — it doesn't have enough RAM to complete after installation. This can be solved by adding some swap memory. However, since the XFS root filesystem takes up all the available space and can't easily be shrunk, only file-based swap is an option. Alternatively, installingzram-toolsand enabling ZRAM works as well. Maybe consider including a small swap partition in the template to make it more functional out of the box.It's great that you offer an alternative way to do custom OS installs via the included netboot.xyz ISO image (it's the only ISO currently available for mounting from the CP, for anyone wondering). It's really cool to see this on a low-end VPS and it opens up almost unlimited customization options. However, because your network doesn't have DHCP and the VPS must be statically configured,
netboot.xyzwon't boot on its own. You have to use VNC and press the 'm' key during startup to access the "failover menu", where you can manually enter static network settings. And when booting into the next image, you'll have to configure the network manually again, asnetboot.xyzdoesn't pass its network config through.I've started preparing a Debian preseed configuration for automated installs. I try to script anything that might need to be done more than once — I don't want to install manually each time. But right now, it requires manually entering network settings at least twice: once during the
netboot.xyzstartup, and again when the OS installer boots (or passing them as kernel options via anothernetboot.xyzmenu). You can embed network settings in the Debian preseed file, but the installer still needs a working network connection to download the preseed file in the first place.I'm not sure how much control VirtFusion gives you over network configuration, but if possible, maybe consider running a DHCP server to hand out network configs for the VPSes. Or maybe automatically passing the network config to
netboot.xyzvia boot options — I'm not sure if it supports this when booting from ISO, but the iPXE boot method does: https://netboot.xyz/docs/booting/ipxeAnyway, this isn't a big deal since these things can be worked around, so it's definitely not a blocker. I'm totally fine doing some "homework" to handle this at this price point. I just decided to take a bit of time to write this comment in case you're open to hearing what might be improved.
Thanks again for the offer! It's great to have an opportunity to try out a new provider and locations nearly for free.
I couldn't resist the 250 GB storage for $4.50 for a year! I still love your TOS, it's no-nonsense, no "legal talk", just plain and simple: "behave, you're on your own"
We don't have any servers lined up right now, but I'm sure we will get one in the near future.
Private VPN and TOR are allowed.
TOR exit nodes, Open proxies/VPN's are NOT allowed.
Thank you for this feedback! Before this, I hadn't had any feedback, good or bad! So thank you, I was impressed with VirtFusion from the gate, but I can't test everything to make sure everything is perfect, which I knew it wouldn't. I will be taking your feedback and looking to implement some of your suggestions.
Thank you again!
Some feedback:
After I install Ubuntu 22, VirtFusion shows CPU is at 100%, disk reading 20 MB/s continuously, memory almost full. SSH doesn't give any output and doesn't let me enter the password I received by email. My guess: 384 MB RAM isn't enough for this.
Ubuntu 24: CPU at 100%, reading 400 MB/s. SSH gets to the password-part, but the password I received by email is incorrect. Reset password failed.
Ubuntu 20: CPU load, disk reading low. Password received by email doesn't work. Reset password failed.
CentOS 9: SSH gives no response after entering the password. CPU 100%, memory full.
CentOS 7: Everything works as expected (but I know nothing about CentOS).
Ubuntu 20 with SSH key: works.
Ubuntu 22 with SSH key: ssh works, CPU 100%, memory full. The system is basically too slow to run
top.AlmaLinux 9: SSH before I can enter password: kex_exchange_identification: read: Connection reset by peer
AlmaLinux 9 with SSH key: works.
I'm back to Ubuntu 20 now, with SSH key. The non-working emailed password may need some TLC. It would be even better if OS-options that can't run in 384 MB RAM are removed.
Thankyou!
I think that is doable within VirtFusion, 384MB is definitely the lowest of the low when it comes to compatibility with running just the bare OS.
Thanks again for your feedback, I will add this to my future changes!
It's pretty cool from a customer perspective as well.
I've looked through their docs and found out that it has support for DHCP servers integrated: https://docs.virtfusion.com/guides/dhcp
A blind guess that's because ssh doesn't allows root logins with password by default. But it's still possible to use root password for "local" logins via VNC feature. But it's not secure, as VNC connection are unencrypted. You can enable password-based root login in
/etc/ssh/sshd_configby settingPermitRootLogin yes, restart ssh or reboot after. If you've ever enter root password on VNC, I suggest changing it through SSH shortly after. It's insecure and possibly leaked after being used on unencrypted plain channel. Passwords entered through VNC are transmitted as plain keystrokes, unlike SSH.Anyways, it's 2025, SSH keys are way better and should be used instead. One real security breach I've seen - attacker has gained remote access (with sudo root permissions) to one "insignificant" border host. While the host itself didn't had any valuable data, it allowed password-based logins and ssh keys weren't mandatory policy. So he replaced sshd with modified version. It was based on the same OpenSSH source and generally behaved like normal, allowing legit users to login as usual. A simple modification just captured every login and password entered and sent it to the hacker on request. While SSH has network channel encrypted, sshd daemon itself decrypts and knows plain password when it's used. It went undetected for a while, until one of users with global admin permissions logged onto that host. Attacker gained his password and used it to breach further into network, on other hosts with sensible data now. This didn't happened with keys, because private keys are never transmitted to the server. Private keys are used only to "sign" some one-time "ticket", proving that client has correct private key, but capturing this handshake won't allow gaining access to other hosts allowing same private key.
As for the low RAM - you can try enabling some swap if you can login at least through VNC after initial OS install:
On ubuntu you can try installing
zram-toolsinstead. It creates swap space from RAM, but have it compressed with fast algo, thus effectively increasing total memory.I'm running a few KVM VPSes with 256 Mb RAM with Debian 12 (but installer for x86_64 arch will fail to boot, had to install Debian 11 -> upgrade instead, or stick to i686 arch alternatively). Minimalistic OS install uses less than 100 Mb of RAM and this can be even lower with additional tweaks. Ubuntu usually comes with pretty much RAM-consuming garbage like unattended install, haveged (?), etc. Those aren't required and should be disabled on lowend hosts with low RAM. Oh and there are dozens of useless (on VM) kernel modules loaded as well afair. Debian offers "cloud" series of kernels, with a narrowed-down set of modules better suitable for VMs. I think Ubuntu should have something similar or you can always manually disable unwanted modules.
I'm pretty sure that CentOS / Alma 9 will be hard to use if possible at possible at all even with swap added. Recent versions of
dnfrequire a lot of RAM. But recent Debian / Ubuntu versions should be pretty fine even with 256 Mb RAM after some tweaking. One of the reasons I prefer Debian over Ubuntu - it's minimal install are minimal indeed and doesn't comes with RAM-hungry garbage.I purchased a 1vCPU, 256MB RAM server and installed Ubuntu Server 24.04 LTS. When I logged into SSH and ran "apt update -y && apt install -y curl && apt install -y socat" to update the server, it got stuck. It was stuck for a long time. After that, I can no longer log into SSH, and pinging my IP address also fails. However, I can still log into the Control Panel and rebuild the system. It seems that my IP address has been blocked. Is there a solution to this problem?
Is internet communication censored in your country? (e.g., China's Great Firewall)
I can successfully ping 8.8.8.8 with no packet loss，but ping my ipv4 address fails. SSH access was working fine when I initially purchased the hosting, but now I'm completely unable to log in. It appears my IPv4 address has been blocked. Are there any solutions?
@Cam
In this case, please open a ticket so I can take a look.
Thanks
@Cam I opened a ticket, and the issue is now resolved. Thanks for your help; I appreciate your service and would recommend it.
Great! Thank you!
Kvm architecture reinstallation script: https://github.com/bin456789/reinstall
OVZ or LXC architecture reinstallation script: https://github.com/LloydAsp/OsMutation
@Cam
It seems code 2025-90OFF-KVMNAT can't work?
What is the renewal price?
Promo is expired.