You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
auto_network allocates IP addresses to machines as they are booted, which interacts poorly with vagrant-hosts. When machines are created and provisioned vagrant-hosts populates the hosts file with all present IP addresses, so machines that are booted later are not added to the hosts file. This means that people either have to provision all machines after the initial creation runs, or defer provisioning till after all machines are created.
The text was updated successfully, but these errors were encountered:
A tractable approach to this problem may be to allocate an IP as soon as an auto-networked machine appears in the Vagrant config rather than allocating on up/destroy. The downside is that the IP pools will begin to accumulate garbage allocations --- so some sort of garbage collection will be necessary. Perhaps a time to live (hey! we just re-wrote DHCP!).
Vagrant 1.6 may enable a different approach since it globally tracks created machines. Under 1.6 and newer, we could allow the pool to fill up and do a one-time rescue of the PoolExhausted exception. When rescued, garbage collect any IP allocation for machines that do not exist in the global Vagrant index and are not defined in the current Vagrantfile.
auto_network allocates IP addresses to machines as they are booted, which interacts poorly with vagrant-hosts. When machines are created and provisioned vagrant-hosts populates the hosts file with all present IP addresses, so machines that are booted later are not added to the hosts file. This means that people either have to provision all machines after the initial creation runs, or defer provisioning till after all machines are created.
The text was updated successfully, but these errors were encountered: