guest network problems after suspend/resume host

[Cecualdo]

Hi all,

I'm running Vista Ultimate x64 as host, VM Workstation 6.04.93057, some 32bit installations as VMs all with bridged networking (dhcp by router). If I put the host into power saving mode (S3 / STR) while a VM is runnig, it looses (changes???) its network connection after resuming the host. From within the VM its not longer possible to access other machines then host. If i start "ipconfig /release" and "ipconfig /renew" it lasts a long time, but i get a new IP. The same if i use VMIP.exe -release -renew, the only host i can reach is the host which runs vmware, no other machine in the same address-range (and no internet). Sounds like after suspend/resume the host, vmware-network changes from "bridged" to "host-only" ??!? Sometimes if I manually disconnect and reconnect the virtual network card to the vm it solves the problem, but usuallly I must close the Workstation window and restart the vm to get proper bridged networking again. I know for sure this works for a while without any problems, but a few days ago it suddenly didn't work any longer. I think that I have not changed any network/vmware settings. vmnet0 is bridged to the first ethernet (second and all other are disabled!). No network difficulties from within the host! I hope that I have clearly stated the problem, should I provide any log/configs (but all looks the same before and after manually restart vmworkstation!) ?

Any ideas how to solve the problem or any hints on how to "debug" vmnet0?

Thanks in advance....

[Peter_vm]

I would start with changing guests to static IP addresses first.

If that does not work, maybe you can disable host suspend/resume function?

[Cecualdo]

Hi, thanks for helping!

I just tried to use NAT for a VM. After S3 and resume host, the network survives (Internet is still reachable). Then I read your post and changed config again to bridged (vmnet0). IP to 192.168.1.220 Gateway, DNS etc. 192.168.1.1 (IPCOP on real hardware). Test: can ping myself, vmwarehost (192.168.1.245), ping to www, all looks fine... suspend host, wait... resume host... no network in vm. I can only ping myself, no other machine, neither internet nor 192.168.1.xxx. But this was working a few weeks ago, and a long time on another machine with similar config.

I don't want to disable S3 on the host, because its a very nice feature to have a fully started workstation with a couple of vms available a few seconds after power-on!

Are there any possibilities like "net start vmnet0"?

[Peter_vm]

You can script disable-enable sequence of your physical NIC on host. Then from a desktop create a shortcut to that script.

[Cecualdo]

If I disable and re-enable the network card of the host after waking up from S3, bridged network in guests will be fully functionally immediately. But network access from within the host works directly without dis/reenable, so I think its a problem with the bindings of host NIC to the bridging protocol respectively vmnet0. Are there any energy saving options for vmnet0/bridging driver? Is it possible to reinstall only the vmnet0-part of Workstation installation? Any ideas on what to look for in vmware.log? Any occurances of "NET" didn't show any helpful hints!

[Cecualdo]

Just now I've started the vm with option to gather full debugging information. Suspend/Resume host, try to browse internet in vm and it worked! Changed nothing, doing again the suspend/resume cycle, networking is functionable a few seconds, then the same behavior as described in above posts. The additional information in vmware.log dos not show any useful hints (for me).

Are there any documentation available for the VMnetBridge entries in the registry? Any ideas on how to 'degub' "system32\DRIVERS\vmnetbridge.sys"?

Sorry for beeing a nuisance...

[Cecualdo]

Reinstall of host-NIC-drivers didn't solve the problem as well as reinstall the complete VMware Workstation.

I'm at a loss what to do!

[Peter_vm]

What about:

You can script disable-enable sequence of your physical NIC on host. Then from a desktop create a shortcut to that script.

Did you even consider that option?

[Cecualdo]

Hi Peter,

the option you consider will work, but I think its a very ugly workaround, because host suspending with powered-on vms worked for a while, without manually disable-enable. Apparently I have damaged my hosts configuration with something I can't remember. Due to the fact that only vms are affected, I thought the error relates to nic/virtual nic connection/mappings.

[Cecualdo]

I've just find out that my nic driver caused the problem (RTL8168B onboard). The latest driver (6.206.0502.2008) will show this behavior, also some previous versions (f.e. the .125.2008). Reproducable an another machine with same nic! The driver shipped with windows update (6.200.1130.2007) will serve my vmnet0 again after resume! Sorry for accusing vmware-bridge to be the cause of that problem.