avsh ("Augmented Vagrant sSH") is a standalone script that can be used in place
vagrant ssh. It provides greatly increased performance and several extra
SSH multiplexing avsh automatically establishes an SSH control socket the first time it's run, which speeds up all subsequent connections by over an order of magnitude.
$ time vagrant ssh -c 'hostname' vagrant-ubuntu-trusty-64 Connection to 127.0.0.1 closed. real 0m2.786s user 0m1.670s sys 0m0.545s $ time avsh hostname vagrant-ubuntu-trusty-64 Shared connection to 127.0.0.1 closed. real 0m0.087s user 0m0.034s sys 0m0.013s
Automatic synced folder switching avsh detects when you're working in a synced folder, and automatically switches to the corresponding directory on the guest before executing commands or starting a login shell.
# config.vm.synced_folder '/home/masonm/asci/, '/var/www/jci', $ pwd /home/masonm/asci/content $ avsh pwd /var/www/jci/content
Run commands on multiple machines If you have a multi-machine setup, you can use
avsh -mto run a command on multiple machines.
$ avsh -m 'openbsd,debian' uname OpenBSD Linux $ avsh -m '/(free|open)bsd/' uname OpenBSD FreeBSD
avsh makes a number of assumptions and shortcuts in order to achieve its performance goals, so it might not work (or be appropriate) for your setup.
- Vagrantfiles are evaluated inside a fake Vagrant environment,
which may cause issues with complex Vagrantfiles that use the non-DSL parts of
Vagrant. Specifically, the
Vagrant.has_plugin?method always returns true, and other methods on the
Vagrantmodule are stubbed out.
- The host must be Linux with OpenSSH 5.6+ or OS X 10.7+. It'll probably work on other Unices, but hasn't been tested.
- No merging of multiple Vagrantfiles.
Download the latest release and put it in your PATH, or run the following to put it in /usr/local/bin/:
curl -sL https://github.com/MasonM/avsh/releases/download/0.1.0/avsh \ | sudo tee /usr/local/bin/avsh > /dev/null \ && sudo chmod ugo+rx /usr/local/bin/avsh
avsh uses the same logic as Vagrant
to find your Vagrantfile. If you only use a single Vagrantfile, add
export VAGRANT_CWD="/vagrantfile_dir/" to your .bashrc or .zshrc to ensure avsh and
Vagrant can always find it.
Usage: avsh [options] [--] COMMAND execute given command via SSH or: avsh [options] start a login shell Options: -m, --machine MACHINE Target Vagrant machine(s). Can be specified as a plain string for a single machine, a comma-separated list for multiple machines, or a regular expression in the form /search/ for one or more machines. If not given, will infer from the Vagrantfile. -r, --reconnect Closes SSH multiplex socket if present and re-initializes it -s, --ssh-options OPTS Additional options to pass to SSH, e.g. "-a -6" -d, --debug Verbosely print debugging info to STDOUT -v, --version Display version -h, --help Display help -c, --command COMMAND Command to execute (only for compatibility with Vagrant SSH)
-m flag is not given for a multi-machine environment,
avsh will try to infer the machine to connect to using the synced folders
defined in your Vagrantfile. It does this by matching the current working
directory against each machine in the order they are defined, using the first
machine that has a matching synced folder (taking into account ancestor
directories). If none are found to match, it will use the primary machine
if one exists, else it falls back to the first defined machine.
When executing a command on multiple machines, automatic synced folder switching is disabled, since that can lead to hard-to-predict behavior. Additionally, a pseudo-TTY will not be allocated (i.e. SSH will not be passed the '-t' flag).