NAT (shared networking) very flaky

S

SorenK2

Bit poster
This has been going on for a while, so it might coincide with the upgrade to macOS Sonoma.

Bridged networking seems to work reliably, but I need shared networking in order to use the host's VPN. Shared networking randomly either seems to fail altogether (can't even ping), or it partially fails (I can ping a host, but running `nmap` claims that TCP ports are closed).

Sometimes, toggling between shared network, off, and back seems to fix it. Sometimes, I need to relaunch Parallels altogether.
 
SorenK2 said:
This has been going on for a while, so it might coincide with the upgrade to macOS Sonoma.

Bridged networking seems to work reliably, but I need shared networking in order to use the host's VPN. Shared networking randomly either seems to fail altogether (can't even ping), or it partially fails (I can ping a host, but running `nmap` claims that TCP ports are closed).

Sometimes, toggling between shared network, off, and back seems to fix it. Sometimes, I need to relaunch Parallels altogether.
Click to expand...
Hello!
Thanks for bringing it up!
Does it happen with a specific VM or with all VMs on your host? Can you collect a report at the moment of reproduction and share it with us, please?
To collect the report click the Parallels icon > Help > Send technical data > check "Attach screenshots..." > press Send Report. Please send the report's 9-digit ID in your reply.
 
Hi,

I'm not sure about other VMs; I almost always use this one. But I can, if it helps, set up a fresh one and see if the problem occurs there.

Having said that: the log ID is 442058018. Here's what I did the five minutes before that log:

1. switched from bridged to NAT networking
2. pinged a host; it times out
3. suspended and resumed the VM; this did not help
4. quit Parallels entirely, waited for prl_* processes to exit
5. launched Parallels again and resumed the VM
6. now, the ping (which resumes) starts working

I guess I should mention this is with a Windows 11 insider build 23580.1000.
 
Hello.

If you could verify the behaviour on a fresh Windows, preferably from express installation, it would help verify if the scope of the issue is applied to either a specific virtual machine or its build, which could potentially could narrow down the cause.

That said, if it's possible, we would request that you attempt to replicate the issue in a fresh virtual machine, and if it is reproducible share a technical report from the new machine as well: To collect the report click the Parallels icon > Help > Send technical data > check "Attach screenshots..." > press Send Report. Please send the report's 9-digit ID in your reply message.
 
Hello Maxim,

the behavior just happened on a fresh Windows. See report 442274318.

In the screenshot in that log, you can see
1) with Parallels still running, requests time out
2) then I quit Parallels (the VM suspends, prl* processes exit, etc.), wait, relaunch it
3) then I run the same ping again, and it works

This is with a Windows VM installed using Parallels's express setup. Basically not much more than "Get Windows 11 from Microsoft".

The 10.1.1.106 endpoint being pinged is a host only reachable through the VPN of the host. Therefore, it won't work using bridged networking; I need a NAT through the host to reach it.
 
You must log in or register to reply here.
Back
Top