Reloading the Vagrantfile causes issue with multiple evaluations
where users expect single evaluation. Instead of allowing local
plugin installation to happen prior to command execution, force
halt after installation and the command to be re-run. This will
prevent multiple loads of the Vagrantfile within a single run.
If you have a vagrant box with proc mounted with
proc /proc proc defaults,hidepid=2 0 0
ps output will be limited to owned process
sudo should extend output
Due to the Vagrantfile being loaded prior to plugin loading to determine
project local plugin information the Vagrant.has_plugin? helper will always
return false when the Vagrantfile is first loaded. To prevent this behavior
we can check for plugins in the plugin data files prior to the plugins
being loaded, and after they have been loaded we can fallback to the
original specification based check.
Force the vagrantfile and configuration loader to be rebuilt after
plugins have been loaded to properly allow newly introduced plugin
configurations to be properly supported within the Vagrantfile.
Prior to this commit, if the args key was a string rather than an array
of strings, the `join` command would fail when appending the arguments
to the run command for a given script. This commit updates that by
ensuring the `args` option is an array prior to joining the arguments.
Prior to this commit, the `Util::Powershell.execute_inline`
method didn't properly join the passed in command and instead attempted
to execute the array of strings. This commit updates that
behavior to join the command array prior to inserting it into the full
powershell command.
When a guest is created, the box metadata information is stored in the
machine data directory. This allows modifications to happen to the
Vagrantfile definition of the box in use (box name change, box version
change, etc) while still allowing the Machine instance of an active
guest successfully load the box currently backing it.
If starting a process while running from within AppImage adjust
the LD_LIBRARY_PATH of the subprocess when the executable exists
outside of the AppImage. This prevents issues of invalid dynamic
library lookups when the AppImage contains common named libraries.
Prior to this commit, the run trigger option wouldn't catch for failures
outside of the #Subprocess.execute raising exceptions. This commit fixes
that by inspecting the exit code result of the subprocess and using the
new `exit_codes` option to determine how to move forward with the
trigger.
`vagrant up` may hang at the "Configuring and enabling network
interfaces..." step when private networks and PTY allocation for SSH
are used.
The newer version of `nmcli` that is part of CentOS now will open a
pager (i.e. `less`) for certain commands if it finds a tty. This
causes the invocations of `nmcli` in `guest_inspection.rb` to hang.
`nmcli` disables the use of a pager in 'terse' (`-t`) output mode,
while still returning enough information for the uses in
`guest_inspection.rb`.
Prior to this commit, the hyper-v provider called an action on a machine
that hadn't fully finished initializing. This commit fixes that by
moving up the initialization of the triggers object next to the rest of
the instance variables of the machine object.
Prior to this commit, vagrant would halt if it could not parse the
metadata file for a box to check for updates with. This commit changes
that to behave like when vagrant fails to download a metadata file and
continue on but warn the user it failed to check for box updates.
Checks only if the Hyper-V hypervisor is enabled instead of all
Hyper-V features. This will allow detection where Hyper-V is in
use for things like ApplicationGuard.
Uses both Get-WindowsOptionalFeature and Get-WindowsFeature to
support check on server and non-server version.
This commit adds a bit more helpful messaging for the user when asking
for a provider that might not exist or does not follow the correct
casing. If it can find a match on a provider it knows about, it will
suggest it.