[EvanK]

With an existing local vm, and existing remote vm, reprovisioned after 1.2.0, evolve:down:files gets an rsync error with host verification:

mcheck@MRTEENY:bam.com (2.2.1@evolve)(master) $ cap staging evolve:down:files
DEBUG[d795296d] Running /usr/bin/env [ -d /var/www/bam.com/staging/master/current/web/wp-content/uploads ] on staging.bam.com
DEBUG[d795296d] Command: [ -d /var/www/bam.com/staging/master/current/web/wp-content/uploads ]
DEBUG[d795296d] Finished in 1.829 seconds with exit status 0 (successful).
INFO[7b4d776e] Running /usr/bin/env vagrant up on localhost
DEBUG[7b4d776e] Command: /usr/bin/env vagrant up
DEBUG[7b4d776e]     Bringing machine 'local' up with 'virtualbox' provider...
DEBUG[7b4d776e]     ==> local: Checking if box 'ubuntu/trusty64' is up to date...
DEBUG[7b4d776e]     ==> local: A newer version of the box 'ubuntu/trusty64' is available! You currently
DEBUG[7b4d776e]     ==> local: have version '14.04'. The latest is version '20150609.0.10'. Run
DEBUG[7b4d776e]     ==> local: `vagrant box update` to update.
DEBUG[7b4d776e]     ==> local: VirtualBox VM is already running.
DEBUG[7b4d776e]     ==> local: Checking for host entries
INFO[7b4d776e] Finished in 5.134 seconds with exit status 0 (successful).
INFO[ff24e5aa] Running /usr/bin/env chmod 600 ./lib/ansible/files/ssh/id_rsa on localhost
DEBUG[ff24e5aa] Command: chmod 600 ./lib/ansible/files/ssh/id_rsa
INFO[ff24e5aa] Finished in 0.004 seconds with exit status 0 (successful).
INFO[6cc3cabc] Running /usr/bin/env vagrant ssh local -c 'cd /vagrant && rsync -e "ssh -i ./lib/ansible/files/ssh/id_rsa" -avvru --delete --copy-links --progress deploy@staging.bam.com:/var/www/bam.com/staging/master/current/web/wp-content/uploads/ /vagrant/web/wp-content/uploads/' on localhost
DEBUG[6cc3cabc] Command: /usr/bin/env vagrant ssh local -c 'cd /vagrant && rsync -e "ssh -i ./lib/ansible/files/ssh/id_rsa" -avvru --delete --copy-links --progress deploy@staging.bam.com:/var/www/bam.com/staging/master/current/web/wp-content/uploads/ /vagrant/web/wp-content/uploads/'
vagrant@127.0.0.1's password: 
DEBUG[1bbb7486] Running /usr/bin/env [ -f /var/log/evolution/wordpress.log ] && [ $(stat -c %U /var/log/evolution/wordpress.log) == 'deploy' ] on staging.bam.com
DEBUG[1bbb7486] Command: [ -f /var/log/evolution/wordpress.log ] && [ $(stat -c %U /var/log/evolution/wordpress.log) == 'deploy' ]
DEBUG[1bbb7486] Finished in 0.077 seconds with exit status 0 (successful).
INFO[0c482bb5] Running /usr/bin/env echo Wed\ Jul\ 29\ 08:23:19\ 2015\  mcheck@MRTEENY.local\   evolve:files:down\  Failure >> /var/log/evolution/wordpress.log on staging.bam.com
DEBUG[0c482bb5] Command: /usr/bin/env echo Wed\ Jul\ 29\ 08:23:19\ 2015\    mcheck@MRTEENY.local\   evolve:files:down\  Failure >> /var/log/evolution/wordpress.log
INFO[0c482bb5] Finished in 0.078 seconds with exit status 0 (successful).
cap aborted!
SSHKit::Runner::ExecuteError: Exception while executing on host staging.bam.com: vagrant exit status: 3072
vagrant stdout: opening connection using: ssh -i ./lib/ansible/files/ssh/id_rsa -l deploy staging.bam.com rsync --server --sender -vvulLogDtpre.iLs . /var/www/bam.com/staging/master/current/web/wp-content/uploads/  (12 args)
vagrant stderr: Host key verification failed.
rsync: connection unexpectedly closed (0 bytes received so far) [Receiver]
rsync error: error in rsync protocol data stream (code 12) at io.c(226) [Receiver=3.1.0]
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/command.rb:97:in `exit_status='
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:56:in `block in _execute'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:37:in `tap'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:37:in `_execute'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:26:in `execute'
/Users/mcheck/clients/bam/bam.com/bower_components/evolution-wordpress/lib/capistrano/tasks/files.rake:28:in `block (5 levels) in <top (required)>'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:14:in `instance_exec'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:14:in `run'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/dsl.rb:10:in `run_locally'
/Users/mcheck/clients/bam/bam.com/bower_components/evolution-wordpress/lib/capistrano/tasks/files.rake:27:in `block (4 levels) in <top (required)>'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/netssh.rb:54:in `instance_exec'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/netssh.rb:54:in `run'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/runners/parallel.rb:13:in `block (2 levels) in execute'
SSHKit::Command::Failed: vagrant exit status: 3072
vagrant stdout: opening connection using: ssh -i ./lib/ansible/files/ssh/id_rsa -l deploy staging.bam.com rsync --server --sender -vvulLogDtpre.iLs . /var/www/bam.com/staging/master/current/web/wp-content/uploads/  (12 args)
vagrant stderr: Host key verification failed.
rsync: connection unexpectedly closed (0 bytes received so far) [Receiver]
rsync error: error in rsync protocol data stream (code 12) at io.c(226) [Receiver=3.1.0]
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/command.rb:97:in `exit_status='
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:56:in `block in _execute'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:37:in `tap'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:37:in `_execute'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:26:in `execute'
/Users/mcheck/clients/bam/bam.com/bower_components/evolution-wordpress/lib/capistrano/tasks/files.rake:28:in `block (5 levels) in <top (required)>'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:14:in `instance_exec'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/local.rb:14:in `run'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/dsl.rb:10:in `run_locally'
/Users/mcheck/clients/bam/bam.com/bower_components/evolution-wordpress/lib/capistrano/tasks/files.rake:27:in `block (4 levels) in <top (required)>'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/netssh.rb:54:in `instance_exec'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/backends/netssh.rb:54:in `run'
/Users/mcheck/.rvm/gems/ruby-2.2.1@evolve/gems/sshkit-1.5.1/lib/sshkit/runners/parallel.rb:13:in `block (2 levels) in execute'
Tasks: TOP => evolve:files:down
(See full trace by running task with --trace)

I can run the command locally, and approve the host key fingerprint:

vagrant ssh local -c 'cd /vagrant && rsync -e "ssh -i ./lib/ansible/files/ssh/id_rsa" -avvru --delete --copy-links --progress deploy@staging.bam.com:/var/www/bam.com/staging/master/current/web/wp-content/uploads/ /vagrant/web/wp-content/uploads/'

Is this due to taking out the ssh configuration from Vagrantfile?

[EvanK]

I can replicate this locally within the test suite, so it's strange that the Travis builds are passing.

Good news is, I'm pretty certain we can work around it by turning off StrictHostKeyChecking in ssh. Bad news is, I'm not totally sure (yet) why this is happening

[EvanK]

Okay, I understand it now. Previous to #59, we were sharing the host's ssh folder (including known hosts) with the vagrant vm, which means that any machine the host has previously connected to was also known to the vm.

Since that shared ssh folder was removed, the vagrant vm is trying to rsync (via ssh) to a machine it does not recognize, hence the prompt.

Rather than turn off strict host key checking, it may be better to set up proper host keys for the remote, via ssh-keygen and/or ssh-keyscan. Looking into this now...

[EvanK]

Since all stages use the same keypair, we should add to (but not overwrite if exists) a known_hosts file during provisioning -- Ansible even has a known_hosts module, but it's only as of 1.9, and I'm not comfortable making that a requirement just yet.

In the meantime, I'm fixing the current issue by turning off StrictHostKeyChecking ¯_(ツ)_/¯

[EvanK]

@mcheck Fixed as of 1.2.1