'vagrant halt' with Debian box takes 30 seconds for Parallels provider

Hello,

I am running Vagrant 1.8.1 with the Parallels provider (1.6.1) on Mac OSX El Capitan, and anytime I 'vagrant halt' a Debian machine, it just hangs for a very long time, even though the machine is actually shut down.

You can verify this by using any standard Vagrantfile and the 'parallels/debian-8.2' box on Atlas--just do a 'vagrant up' followed by a 'vagrant halt'. It looks to me like the halt command is never returning back to the caller, so it just waits for the 30 second timeout.

Is there a way I can work around this problem?

 

I guess you mean 30 minutes, not seconds, right?
Yes, that's the known issue of Vagrant. It was reported to our provider (issue#240), but after investigation I've got that it should be fixed in the core Vagrant:
https://github.com/mitchellh/vagrant/issues/6207#issuecomment-168972413

I'm really sorry, but seems like this issue could not be worked around. Vagrant just hangs sending SSH keep-alive messages and you can only interrupt the process by pressing Ctrl+C twice.

P.s. This bug affects only Debian Linux guests. Other Linux guests, such as CentOS and Ubuntu could be halted without issues.