High network latency on vm

hello
I’m new to OpenNebula,i installed OpenNebula(4.12.1) cluster with 4 nodes.Today i found some vm really slow. So i try to ping the virtual machine from the server(vm on this),here is the result:

And below is some info about this vm

top - 13:46:26 up 6 days, 21:04, 1 user, load average: 0.01, 0.03, 0.05
Tasks: 71 total, 2 running, 69 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.3 us, 0.3 sy, 0.0 ni, 98.7 id, 0.7 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem: 2050048 total, 1648272 used, 401776 free, 68852 buffers
KiB Swap: 0 total, 0 used, 0 free. 914500 cached Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
8488 root 20 0 23536 1500 1076 R 0.3 0.1 0:01.61 top
1 root 20 0 33512 2796 1444 S 0.0 0.1 0:00.75 init
2 root 20 0 0 0 0 S 0.0 0.0 0:00.03 kthreadd
3 root 20 0 0 0 0 S 0.0 0.0 0:03.71 ksoftirqd/0
5 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0H

root@ip-192-168-30-25:/home/mpr# free -m
total used free shared buffers cached
Mem: 2002 1609 392 0 67 893
-/+ buffers/cache: 649 1352
Swap: 0 0 0
root@ip-192-168-30-25:/home/mpr#

Thanks for reading.Your help will be appreciated

Did you find a solution to your problem. I’m experiencing the same in a new cluster with last version of opennebula. I’m using dummi bridge networking and local storage.

Thanks

Hi @rubentolosa,

Could you provide some more details? Like OpenNebula version, number of hosts, network topology, VM’s guest OS, VM’s configuration, especially the VM’s NIC model used, etc.

Best Regards,
Anton Todorov

Sorry for disturbing, I found the root of the problem, but will post details to help others:

OpenNebula 5.6.1 on top of Ubuntu 18.04
1 controller node + 4 compute nodes (Dell R740XD) All connected via 10 Gb Ethernet BaseT to a 10 Gb Eth Switch.
Storage is local 4 to 6 SSD disks in raid 10 (ZFS) in each compute node.

For creating virtual machines I’ve used de ubuntu 18.04 - KVM image from the OpenNebula public market. And also cloned that template to use it with several disk images imported from other virtualization clusters i had.

The ping latency problem (+ than 1 second in some responses) was specially obvious with imported images that used ubuntu 12.04, but with 16.04 imported machines and even new instances from the template for 18.04 the network was slower than normal ( less than 1 Gbps )

I didn’t notice that the template from the market did not set a NIC model by default, the field was empty, so when using it to create a vm it was booted with the default for KVM (realtek rtl8139).

I’ve edited the template to use virtio as default NIC model, detached and attached a virtio NIC in all running instances, and not only the ping problems disapeared, but now all my VM can communicate at, at least, 10 Gbps.

It was just a stupid error on my side… I hope my experience can help others.

Thanks for your time…