0

When I run top from my VPS, at the %Cpu(s) row, I see a few values, e.g. %us, %sys, %wa, etc

I'd like to ask whether these are the stats for the VPS only, or for the whole node (dedicated server)?

For example, let's say %sys = 10, does it mean that the Percentage of the CPU for system processes of the node is 10%, or does it mean Percentage of the CPU for system processes within my VPS is 10% of the allocated CPU of the VPS (Since system processes may refer to either the dedicated server or only the VPS)

Let's take another example, %us = 20, does it mean the sum of Percentage of the CPU for user processes of all VPS servers in the node is 20%, or does it mean the sum of Percentage of the CPU for user processes of all users within my VPS is 20% of the allocated CPU of the VPS (since user processes may refer to either users within my VPS, or the VPS users within the node)?

I'm not clear about this, because VPS is also a system and it has its own users, and the node is also a system and it has its own users (the users of the node are the VPSs, if I'm guess correctly).

Are there any difference in explanation between OpenVZ or KVM VPS?

Thank you very much

aye
  • 229
  • 3
  • 11
  • Requests for product, service, or learning material recommendations are off-topic because they attract low quality, opinionated and spam answers, and the answers become obsolete quickly. Instead, describe the business problem you are working on, the research you have done, and the steps taken so far to solve it. – TomTom Oct 18 '17 at 07:15
  • So do you mean VPS is a product and we should not ask questions about that? If you read the question you will see I asked about the top command explanation, which sounds similar to a few questions on serverfault, except that this is specifically for VPS – aye Oct 18 '17 at 15:46
  • 2
    No, it is more around the line that this is not a place to teach you the basics. If you do not know how to administrate a system, hire someone who knows. You ask EXTREMELY basic questions. – TomTom Oct 18 '17 at 15:53
  • The problem is, I have been managing VPS servers for years and it works most of the time. I don't know my level of knowledge compared to others. Day by day I learn something from here and there. If I know the question is EXTREMELY basic I wouldn't ask, since I don't know. If I know already there is no point to ask. In this question, It is just like other questions got high vote in this site, except that this is in side a VPS, and I haven't seen it any where on the internet so I ask. Next time, can you just straight away delete my questions if you feels it doesn't fit this site? Thanks! – aye Oct 18 '17 at 16:49

1 Answers1

2

The stats reported by your OS are for your VPS only and not for the hypervisor (the physical server that runs the virtualization software and which manages all the different VPS’s)

HBruijn
  • 77,029
  • 24
  • 135
  • 201
  • Thanks for your answer. Actually I feel it is weird because, in general, we always use `%wa` to see if the disk of the node is heavily abused. So from within the VPS, we can kind of know the node disk state. – aye Oct 18 '17 at 15:43
  • 1
    In a VPS high IO wait times are still a bad thing and may be an indication that either your VPS is overloaded and hitting the IO limits set and enforced by the hypervisor (fi to ensure that every customer gets a guaranteed minimal IO level and a single run-away VPS won't impact other customers too much) or that maybe the hypervisor is overloaded and can't provide the IO performance requested by your and all other VPS's combined. The latter requires access to the hypervisor to determine. – HBruijn Oct 18 '17 at 16:19