I can ssh just fine, but ansible says "no route to host" - ssh

Question

I wrote a script to run up several vms using vagrant, which I have to then provision with ansible. Unfortunately my host is a windows machine, so I thought I could solve the issue by putting all the vms into a vpn and then provision them from another machine in the same vpn.
In theory, it works... I can ssh into the other machines without trouble. But when I run my ansible playbook, ansible fails.
At first I got the message "ssh: connect to host 10.1.2.100 [10.1.2.100] port 22: No route to host" when running ansible with -vvvv
This was in the evening, and I was very tired, and this error didn't recur the following morning. Not sure if it's got something to do with the vm I'm doing deployment from being rebooted in the meantime, or the receiving machine being destroyed and uped completely since then. In any case, the problem has not gone away.
results now, after recreating both vms:
# ansible-playbook -i vms -k -u vagrant vms.yml -vvvv
result:
<10.1.2.100> ESTABLISH SSH CONNECTION FOR USER: vagrant <10.1.2.100>
SSH: EXEC sshpass -d14 ssh -C -vvv -o ServerAliveInterval=50 -o
User=vagrant -o ConnectTimeout=10 -tt 10.1.2.100 '( umask 22 && mkdir
-p "$( echo $HOME/.ansible/tmp/ansible-tmp-1455781388.36-25193904947084 )" && echo
"$( echo $HOME/.ansible/tmp/ansible-tmp-1455781388.36-25193904947084
)" )' fatal: [10.1.2.100]: FAILED! => {"failed": true, "msg": "ERROR!
Using a SSH password instead of a key is not possible because Host Key
checking is enabled and sshpass does not support this. Please add
this host's fingerprint to your known_hosts file to manage this
host."}
So far so clear. I ssh into the other instance to add it to the known hosts. This works without any trouble.
Back to ansible, I try the same command again. The result now is:
<10.1.2.100> ESTABLISH SSH CONNECTION FOR USER: vagrant <10.1.2.100>
SSH: EXEC sshpass -d14 ssh -C -vvv -o ServerAliveInterval=50 -o
StrictHostKeyChecking=no -o User=vagrant -o ConnectTimeout=10 -tt
10.1.2.100 '( umask 22 && mkdir -p "$( echo $HOME/.ansible/tmp/ansible-tmp-1455782149.99-271768166468916 )" &&
echo "$( echo
$HOME/.ansible/tmp/ansible-tmp-1455782149.99-271768166468916 )" )'
<10.1.2.100> PUT /tmp/tmpXQKa8Z TO
/home/vagrant/.ansible/tmp/ansible-tmp-1455782149.99-271768166468916/setup
<10.1.2.100> SSH: EXEC sshpass -d14 sftp -b - -C -vvv -o
ServerAliveInterval=50 -o StrictHostKeyChecking=no -o User=vagrant -o
ConnectTimeout=10 '[10.1.2.100]' fatal: [10.1.2.100]: UNREACHABLE! =>
{"changed": false, "msg": "ERROR! SSH Error: data could not be sent to
the remote host. Make sure this host can be reached over ssh",
"unreachable": true}
Well, I made sure the host was reachable by ssh, thank you very much! Ansible still can't get through, and I'm about to get a brain tumor from thinking of things that might be the problem.
Any suggestions what might be the problem?

Answer 1

This issue was reported here, with some workarounds:
https://github.com/ansible/ansible/issues/15321
The consensus seems to be either to a. use ansible_password or b. use -u username in the connection parameters. However, any number of things can disrupt an SSH connection in ways that make it look "unreachable" to higher level apps, so I recommend going through each of the steps outlined in that ticket.