Jump to content
Welcome to our new Citrix community!
  • 0

Vewry Slow VM


Regis Pietro

Question

Hello, we have two different XenServers:

  • SunFire x4450 with 128G Memory, 4x Quad Core Intel Xeon X7350 @2.93GHz
  • SunFire x4440 with64G Memory, 4 x Quad Core AMD Opteron 8356

The Storage disks and network interfaces are the same for both servers.

While we don't have any problem running old O.S. (Windows 7, CentOS6, Unubtu 14/16)  on both Server, if we try to run the same VM with a newer O.S. (Windows 10, Ubuntu 18, CentOS 7.6), the VM runs "normally" on XenServer/Sunfire X4440, but the same VM is very slow on SunFire X4450.

For example, a Windows 10 VM (8G RAM, 4 VCPU) after boot remains with 100% CPU indefinititely.

Is there perhaps some setup (on bios, on xenserver) we can do for resolving this issue?

The attached file shows VM Performance on both machines, the difference is very clear.

 

 

 

XenServer.Performance.Problem.docx

Link to comment

16 answers to this question

Recommended Posts

  • 0

Yes,

the strange thing is that the same VM (we imported the VM on both Server) is very slow on one XenServer (sunfire x4450) and normal on the other server (sunfire x4440).

We have this problem for both Windows and Linux VM, and only for VM with newer O.S.

If we instead import a VM with an older O.S. (for example CentOS6), the VM seems to have the same speed on both XenServer.

 

 

Link to comment
  • 0

Hello again,

we tried network/disk speed test on the same VM on both server, and results are almost the same.

The problem seems not be caused by slow ethernet or disk, also because we suppose we should have the same problem also with VM running older O.S..

 

XenServer runnig un sunfire x4450 (slow speed VM):

root@vubuntu-prova:/tmp# iperf3 -c 192.168.8.18
Connecting to host 192.168.8.18, port 5201
[  4] local 192.168.8.169 port 35158 connected to 192.168.8.18 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   107 MBytes   900 Mbits/sec    0    614 KBytes
[  4]   1.00-2.00   sec   107 MBytes   900 Mbits/sec    0    614 KBytes
[  4]   2.00-3.00   sec   108 MBytes   902 Mbits/sec    0    614 KBytes
[  4]   3.00-4.00   sec   109 MBytes   913 Mbits/sec    0    614 KBytes
[  4]   4.00-5.00   sec   107 MBytes   899 Mbits/sec  110    452 KBytes
[  4]   5.00-6.01   sec   106 MBytes   889 Mbits/sec    0    516 KBytes
[  4]   6.01-7.00   sec   107 MBytes   901 Mbits/sec    0    557 KBytes
[  4]   7.00-8.00   sec   109 MBytes   917 Mbits/sec    0    585 KBytes
[  4]   8.00-9.00   sec   109 MBytes   916 Mbits/sec    0    595 KBytes
[  4]   9.00-10.00  sec   110 MBytes   925 Mbits/sec    0    605 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.06 GBytes   906 Mbits/sec  110             sender
[  4]   0.00-10.00  sec  1.05 GBytes   903 Mbits/sec                  receiver

iperf Done.

root@vubuntu-prova:~# dd if=/dev/zero of=/tmp/dump bs=512 count=1000 oflag=dsync
1000+0 records in
1000+0 records out
512000 bytes (512 kB, 500 KiB) copied, 25.3497 s, 20.2 kB/s
root@vubuntu-prova:~# dd if=/dev/zero of=/tmp/dump bs=1G count=1 oflag=dsync
1+0 records in
1+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 30.8417 s, 34.8 MB/s

 

XenServer runnig un sunfire x4440 (normal speed VM):

 

root@vubuntu-prova:~# iperf3 -c 192.168.8.18
Connecting to host 192.168.8.18, port 5201
[  4] local 192.168.8.169 port 51720 connected to 192.168.8.18 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   114 MBytes   956 Mbits/sec    0    370 KBytes
[  4]   1.00-2.00   sec   112 MBytes   942 Mbits/sec   12    361 KBytes
[  4]   2.00-3.00   sec   112 MBytes   940 Mbits/sec    0    361 KBytes
[  4]   3.00-4.00   sec   112 MBytes   940 Mbits/sec   12    349 KBytes
[  4]   4.00-5.00   sec   112 MBytes   940 Mbits/sec    0    363 KBytes
[  4]   5.00-6.00   sec   112 MBytes   940 Mbits/sec    0    363 KBytes
[  4]   6.00-7.00   sec   112 MBytes   940 Mbits/sec    0    363 KBytes
[  4]   7.00-8.00   sec   112 MBytes   939 Mbits/sec   13    351 KBytes
[  4]   8.00-9.00   sec   112 MBytes   940 Mbits/sec   13    291 KBytes
[  4]   9.00-10.00  sec   112 MBytes   940 Mbits/sec    0    358 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec   50             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.

root@vubuntu-prova:/tmp# dd if=/dev/zero of=/tmp/dump bs=512 count=1000 oflag=dsync
1000+0 records in
1000+0 records out
512000 bytes (512 kB, 500 KiB) copied, 22.3643 s, 22.9 kB/s
root@vubuntu-prova:/tmp# dd if=/dev/zero of=/tmp/dump bs=1G count=1 oflag=dsync
1+0 records in
1+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 33.328 s, 32.2 MB/s

 

 

Link to comment
  • 0

The problem is that when we use the same VM on the two XenServers, the CPU load is very different, and the VM on XenServer Sunfie x4450 is very much slower than the VM on XenServer SunFile x4440.

If we check network or disk performance we don't see differences (is perahps a CPU/memory problem?).

 

 

 

Link to comment
  • 0

Ah, okay. Slowness is usually network related. Since this is CPU related and both CPU's seem to be on the HCL, 

that makes it a lot more difficult. Any chance you have support to contact Citrix directly for an answer ? If not

the Management Agent and nothing in the logs as a guide, I'm not sure what to suggest other than opening

a ticket with Citrix or going through a 3rd party vendor for assistance.

 

--Alan--

 

 

Link to comment
  • 0

Did you check the dom0 memory and if needed, increase it such that it doesn't run out of resources? Run "top" to make sure the CPU isn't maxed out on your XenServer host and also that ou are not swapping a lot, if at all.  You can also run "xentop" to monitor VM (domU) usage for your clients.

 

-=Tobias

Link to comment
  • 0

Hello,

yes we tried to raise dom0 memory to 8G, but without results.

If we run top during VM running we don't see problems on XenServer (CPU load < 0.5, memory used < 600000, no swap), but if we try xentop we see that Windows 10 VM CPU load (VM with 4 CPU) is 400%. If the allocate 8 CPU to VM the CPU load become 800%, but VM doesn't becom faster.

The strange thing is that we have this problem only for VM with newer O.S. (Windows 10, Ubuntu 18): VM with older O.S. (Windows 7, Debian, Centos < 7.6) run without problem.

 

Link to comment
  • 0

I would try to build a new VM using the latest template and see if the same thing happens.  Since it's not happening with other OS versions, I would have to guess that the issue lies with how there specific VMs are created. Under XS 7.1 and 7.6 we do not see anything like that happening.

 

-=Tobias

Link to comment
  • 0

Hello,

we have 2 sunfire x4440 server and 3 sunfire x4450 server, and the behaviour is the same on all server:

  • VM with normal speed on sunfire x4440 (two server)
  • VM with very slow speed on sunfire x4450 (three server)

We tried both new VM and the same VM imported on the XenServers, and the reaults are always the same.

As we already said, we instead don't have any problem with VM with older O.S..

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...