0

I have just setup a new SolusVM Xen Node at my DC. The server is colocated there, it's owned by me. It was in Florida working perfectly as an OpenVZ node two days before. It was taken offline and sent to a DC in Chicago for colo. Two days ago the ram was displaying fine, 35GB.

Once it got to the DC in Chicago I had them re-install CentOS 6 since this was now going to be a Xen Node, so a fresh install was needed.

Once installed I logged in and installed Xen for SolusVM using their install script. Then I ran free -m and got this:

             total       used       free     shared    buffers     cached
Mem:           924        482        442          0         25        297
-/+ buffers/cache:        158        766
Swap:         4093          0       4093

924MB is ram is all it's showing! When it left Florida it had 35GB. So I was a bit concerned but thought maybe it was something Solus did when it installed Xen.

Once I got done setting up the node in SolusVM panel I went into manage the node and lone behold it says it has 35GB of ram:

enter image description here

Since this is my first time installing Xen PV for Solus does Solus and Xen partition off the ram for use for VPS units and system ram separately and that's why it's not showing it in free? Or was something botched by the DC?

If so, why does Xen do this?

jfreak53
  • 188
  • 1
  • 3
  • 25

1 Answers1

2

Memory for the dom0 is probably limited via kernel cmdline. You can check that by looking into:

cat /proc/cmdline

If there is something like: dom0_mem=min:1024M,max:1024M, then that's the reason for unexpected RAM readings.

Jakov Sosic
  • 5,157
  • 3
  • 22
  • 33