FAQ

General

What is a Virtual Environment (Virtual Private Server, VPS, VE)?

See VE

Who needs OpenVZ? How it can be used?

See Use cases

How is OpenVZ different from other technologies?

See Introduction to virtualization

How is OpenVZ updated and why it is secure?

See Security

I want to show my appreciation to OpenVZ and put some logo to my site. Where to get it?

See Artwork

Installation and upgrade

What hardware is supported by OpenVZ kernel?

See Virtuozzo HCL.

Why there are different kernel flavours available and what do they mean?

See Different kernel flavors (UP, SMP, ENTERPRISE, ENTNOSPLIT)

How do I rebuild the kernel?

See Kernel build

What does 021stab018 in OpenVZ kernel version mean?

See Kernel versioning

How can I check package signatures?

See Package signatures

Is it possible to run x86 VPS on a x86_64 arch?

Sure :) We actually did some work on that to enable migration of x86 VE from x86 to x86_64 and back, and to enable using 32-bit iptables in 32bit VE on an x86_64 system.

Networking

How do I set up VPN for a VE?

See VPN via the TUN/TAP device

What is veth and how do I use it?

See Virtual Ethernet device

User Beancounters

How can I reset failcnt in /proc/user_beancounters?

In short — only by stopping and starting a VE. In some cases you need to keep VE in a stopped state for a few minutes.

Long answer: there can be many application who reads /proc/user_beancounters, and thus if you will reset it you gonna have problems with those other apps.

Consider what happens if you will reset you sent/received packets/bytes statistics on the network interface — programs which reads will be screwed up.

So the proper thing to watch for is not the current value of, say, failcnt, but whether it is increased (from the previous reading) or not. You can write a simple shell script to do just that.

Troubleshooting

My kernel crashed. What should I do?

See When you have an oops

I see a lot of processes in D state. What does that mean?

See Processes in D state