Unexpected Behavior When Switching Host Network

[crfroehlich]

When prepping to take my Windows Server 2008 R2 (x64) laptop home from the office, I suspend my VMs, exit VMWare (7.1 build 261024), and put the host to sleep. On resume, either at my home network or essentially any network different from my last known good state, I can get bizarre and unexpected behavior. VMs can take 10-15 minutes to resume. Once resumed, network access is limited and I cannot VPN into my office network. Sometimes performance on the host takes an enormous hit to the point that it is not possible to do anything.

All of my VMs are NAT'd, which I suspect is related to this issue. Occasionally, simply restarting the VMWare NAT service on the host and disabling/reenabling the NICs inside the VMs will restore full network connectivity. Frequently, a complete reboot of the host and restart of individual VMs is required.

Other coworkers are having this issue running Windows 7 (Enterprise x64) and VMWare 7 and 7.1 (We're all using NAT'd VMs).

Any thoughts? Short of shutting down each VM and shutting down the host to facilitate a clean boot on a new network, I'm at a loss.

[nithinr]

hi,

Sorry that you are facing these probelms. Lets try to understand the networking related problems, first. Can you ps. answer these questions:

- What host is this? Windows / Linux?

- w.r.t networking problems, we want to understand why connecting to the

VPN is not working. Does networking work in general once you transition

to the new network? Is it only connecting to the VPN that is failing?

Can you provide us a sniffer trace on vmnet8 when you try to connect to

the VPN? Instructions to run the sniffer:

Windows (Run these commands in the terminal with ADMINISTRATOR)

1) cd to the VMware installation directory. Typically in C:\Program

Files\VMware\VMware Workstation

2) vnetsniffer /w %temp%/vmnet8.pcap vmnet8

Linux:

1) sudo vmnet-sniffer -w /tmp/vmnet8.pcap /dev/vmnet8

Provide the vmnet8.pcap thus created.

[louyo]

PMJI:

>>Any thoughts? Short of shutting down each VM and shutting down the host to facilitate a clean boot on a new network, I'm at a loss.

When you suspend your host all network settings are retained. When you wake it up on a different LAN, it will try to use all settings and caches that it "remembers" from the LAN where it went into limbo. This can be a problem whether you are running VM's or not but would seem like it would be compounded if you suspend the VM's.

If the host is using dynamic addressing, you might try doing: "ipconfig /release" before suspending and "ipconfig/renew" upon resuming. You very well may need to do the same thing on the VM's. Probably not necessary, but when you resume, you can also clear the ARP cache by doing: "ARP -d * " and clear the DNS cache with:"ipconfig /flushdns" in case your home connection uses different DNS servers.

Edit: Simple scripts could do this.

If I bring a different machine on line before resuming a suspended VM, it is not unusual to see an IP address conflict error. ipconfig/release and renew resolves. Most of the time, I use static IP addresses to avoid this but if you are on different subnets, this could be difficult.

Lou