Hello,
We have a setup of few VMs (all with Linux 64bit) connected with "LAN" between themselves, and I get informations, that network performance within this LAN is bad. First I recommended to install VMTools in these VMs and change NIC to vmxnet, what was done, but still they have big lags with network communication. Below is sample ping test from one VM (in the "LAN") to another
4 bytes from 192.168.0.21: icmp_seq=34 ttl=64 time=0.459 ms
64 bytes from 192.168.0.21: icmp_seq=35 ttl=64 time=0.487 ms
64 bytes from 192.168.0.21: icmp_seq=36 ttl=64 time=2574 ms
64 bytes from 192.168.0.21: icmp_seq=37 ttl=64 time=1575 ms
64 bytes from 192.168.0.21: icmp_seq=38 ttl=64 time=575 ms
64 bytes from 192.168.0.21: icmp_seq=39 ttl=64 time=0.422 ms
I don't know linux much, so I don't know, but maybe there is some simple thing to change that will make net performance better? What is the best way to check network bottlenecks in linux VMs (what can cause these problems)?
The VM are on the same ESX, on the same portgroup?
Which kind of distro are you using?
Andre
**if you found this or any other answer useful please consider allocating points for helpful or correct answers
Andre,
Thanks for your reply, I wanted to make the same tests yesterday (I thought about moving problematic VMs to one ESX host), but customer confirmed, that since yesterday no other problems occur (after installing VM tools), so right now I won't bother this anymore, and if later problem occurs, I will then try to analyze this.
Maybe was only caused by a high load of on of the VM.
CPU load (inside the VM) can vary a network performance.
Andre
**if you found this or any other answer useful please consider allocating points for helpful or correct answers
this CPU load inside VM would be seen in VC performance charts? Or should it be monitored by some linux tools inside VM?
this CPU load inside VM would be seen in VC performance charts?
Yes, in VM performance graph.
Or should it be monitored by some linux tools inside VM?
If you want you can also do so.
For example using vmstat.
Andre
**if you found this or any other answer useful please consider allocating points for helpful or correct answers