I have VM Workstation 8.0.2 installed on a host with
– OS as Ubuntu 64 bit 12.04
– 32 CPUs
– 64 GB of RAM
– 4 x Intel NICs each 1Gbps speed
My guest VM is
– 32bit Ubuntu 10.10
– 4GB of RAM
– 4 CPUs
– 2 Virtual NICs at speeds of 1000Mbps (I checked with ethtool)
– VMWare client tools installed (and reinstalled several times)
I get max of 120 Mbps between the guest and the host (measured using many tools; i.e. iperf and cp over NFS4 from and to guest). On the other hand, the performance of the host (where the VM Work Station 8 is installed) to another host on the same network is 900+Mbps. BTW, pinging from the host to the guest is 600ms and to another host on the network around 95 ms. In other words, the VM Guest performance is 1/6 the performance of calling another host on the network.
Most solutions on the net (including the one below),
http://communities.vmware.com/thread/105389
recommended turning OFF tcp offloading, and my output on the host and guest as follows:-
#host, same output for other eth1,2,&3
sudo ethtool –k eth0
Offload parameters for eth0:
rx-checksumming: off
tx-checksumming: off
scatter-gather: off
tcp-segmentation-offload: off
udp-fragmentation-offload: off
generic-segmentation-offload: off
generic-receive-offload: on
large-receive-offload: off
rx-vlan-offload: on
tx-vlan-offload: on
ntuple-filters: off
receive-hashing: on
#Guest
rx-checksumming: on
tx-checksumming: off
scatter-gather: off
tcp-segmentation-offload: off
udp-fragmentation-offload: off
generic-segmentation-offload: off
generic-receive-offload: off
large-receive-offload: off
I looked for solutions on the net and couldn’t find any, any ideas guys?