Centos 5.4 x86_64 Guest 4GB RAM Breaks paravirt network
Posted: 7. May 2010, 17:43
Greeting fellow VritualBox users!
I have several Centos 5.4 x86_64 Guests running in VirtualBox 3.1.6. uname reports identical kernels across all three VMs; each guest has 2CPUs assigned.
Prior to today, I was running each of these guests with <= 2GB RAM. In this configuration paravirtualized network drivers were working properly for one DHCP guest and the other two Static-IP guests.
Today I needed to run several tests which required assigning additional RAM to each of these VMs, so i upped the allocated RAM to 4096MB on each. On reboot, none of the VMs would connect to the network properly. The interfaces would come up (ifup) but no packets were flowing out-from (ssh, ping, nslookup) or in-to (ssh, ping) the guests. I had to shutdown the VMs again and reconfigure them to use the Intel PRO/100 T Server (82543GC), whereupon the Guest networking returned to its nominal state. Has anyone else run into issues with paravirtualized networking and large RAM allocations?
I have several Centos 5.4 x86_64 Guests running in VirtualBox 3.1.6. uname reports identical kernels across all three VMs; each guest has 2CPUs assigned.
Code: Select all
uname -a
Linux uscobrmfa-gg-50 2.6.18-164.15.1.el5 #1 SMP Wed Mar 17 11:30:06 EDT 2010 x86_64 x86_64 x86_64 GNU/LinuxToday I needed to run several tests which required assigning additional RAM to each of these VMs, so i upped the allocated RAM to 4096MB on each. On reboot, none of the VMs would connect to the network properly. The interfaces would come up (ifup) but no packets were flowing out-from (ssh, ping, nslookup) or in-to (ssh, ping) the guests. I had to shutdown the VMs again and reconfigure them to use the Intel PRO/100 T Server (82543GC), whereupon the Guest networking returned to its nominal state. Has anyone else run into issues with paravirtualized networking and large RAM allocations?