ENT-5725: Added vagrant provider for spawning VMs

MANIFEST.in

@@ -1 +1,2 @@
 include cf_remote/nt-discovery.sh
+include cf_remote/Vagrantfile

cf_remote/Vagrantfile

@@ -0,0 +1,64 @@
+# -*- mode: ruby -*-
+# vi: set ft=ruby :
+require 'json'
+
+config_path = File.expand_path("../config.json", __FILE__)
+if !File.file?(config_path)
+  abort "Need a config file for this host at " + config_path
+end
+config = JSON.parse(File.read(config_path))
+
+VM_BOX = config['box']
+VM_COUNT = config['count']
+VM_MEMORY = config['memory']
+VM_CPUS = config['cpus']
+VM_PROVISION = File.expand_path(config['provision'], __FILE__)
+VM_NAME = config['name']
+SYNC_FOLDER = config['sync_folder']
+
+Vagrant.configure("2") do |config|
+
+  (1..VM_COUNT).each do |i|
+
+    config.vm.define "#{VM_NAME}-#{i}" do |node|
+      node.vm.hostname = "#{VM_NAME}-#{i}"
+      node.vm.network "private_network", ip: "192.168.56.#{9 + i}"
+      node.vm.box = VM_BOX
+
+      node.vm.provision "shell" do |s|
+        ssh_pub_key = File.readlines("#{Dir.home}/.ssh/id_rsa.pub").first.strip
+        s.inline = <<-SHELL
+        echo #{ssh_pub_key} >> /home/vagrant/.ssh/authorized_keys
+        echo #{ssh_pub_key} >> /root/.ssh/authorized_keys
+        SHELL
+      end
+      node.vm.provision "bootstrap", type: "shell", path: VM_PROVISION
+      node.vm.synced_folder  ".", "/vagrant",
+                  rsync__args: ["--verbose", "--archive", "--delete", "-z", "--links"]
+      node.vm.synced_folder "#{SYNC_FOLDER}", "/synched_folder",
+                  rsync__args: ["--verbose", "--archive", "--delete", "-z", "--links"]
+
+      # https://bugs.launchpad.net/cloud-images/+bug/1874453
+      NOW = Time.now.strftime("%d.%m.%Y.%H:%M:%S")
+      FILENAME = "serial-debug-%s.log" % NOW
+      node.vm.provider "virtualbox" do |vb|
+          vb.memory = VM_MEMORY
+          vb.cpus = VM_CPUS
+          vb.customize [ "guestproperty", "set", :id, "/VirtualBox/GuestAdd/VBoxService/--timesync-set-threshold", 1000 ]
+          vb.customize [ "modifyvm", :id, "--uart1", "0x3F8", "4" ]
+          vb.customize [ "modifyvm", :id, "--uartmode1", "file",
+          File.join(Dir.pwd, FILENAME) ]
+      end
+
+      node.vm.provider :libvirt do |v, override|
+        v.memory = VM_MEMORY
+        v.cpus = VM_CPUS
+        # Fedora 30+ uses QEMU sessions by default, breaking pretty much all
+        # previously working Vagrantfiles:
+        # https://fedoraproject.org/wiki/Changes/Vagrant_2.2_with_QEMU_Session#Upgrade.2Fcompatibility_impact
+        v.qemu_use_session = false
+        override.vm.synced_folder "#{SYNC_FOLDER}", "/synched_folder", type: :rsync
+      end
+    end
+  end
+end

cf_remote/aramid.py

@@ -29,6 +29,7 @@
 import time
 from urllib.parse import urlparse
 from cf_remote import log
+from cf_remote.paths import SSH_CONFIG_FPATH
 
 DEFAULT_SSH_ARGS = [
     "-oLogLevel=ERROR",
@@ -37,6 +38,8 @@
     "-oBatchMode=yes",
     "-oHostKeyAlgorithms=+ssh-rsa",
     "-oPubkeyAcceptedKeyTypes=+ssh-rsa",
+    "-F",
+    "{}".format(SSH_CONFIG_FPATH),
 ]
 """Default arguments to use with all SSH commands (incl. 'scp' and 'rsync')"""
 
@@ -338,10 +341,11 @@ def execute(
         if host.port != _DEFAULT_SSH_PORT:
             port_args += ["-p", str(host.port)]
         proc = subprocess.Popen(
-            ["ssh", host.login]
+            ["ssh"]
             + DEFAULT_SSH_ARGS
             + port_args
             + host.extra_ssh_args
+            + [host.login]
             + [commands[i]],
             stdout=subprocess.PIPE,
             stderr=subprocess.PIPE,

cf_remote/commands.py

@@ -18,9 +18,11 @@
 from cf_remote.web import download_package
 from cf_remote.paths import (
     cf_remote_dir,
+    cf_remote_packages_dir,
     CLOUD_CONFIG_FPATH,
     CLOUD_STATE_FPATH,
-    cf_remote_packages_dir,
+    SSH_CONFIG_FPATH,
+    SSH_CONFIGS_JSON_FPATH,
 )
 from cf_remote.utils import (
     copy_file,
@@ -37,7 +39,14 @@
     CFRChecksumError,
     CFRUserError,
 )
-from cf_remote.spawn import VM, VMRequest, Providers, AWSCredentials, GCPCredentials
+from cf_remote.spawn import (
+    CloudVM,
+    VMRequest,
+    Providers,
+    AWSCredentials,
+    GCPCredentials,
+    VagrantVM,
+)
 from cf_remote.spawn import spawn_vms, destroy_vms, dump_vms_info, get_cloud_driver
 from cf_remote import log
 from cf_remote import cloud_data
@@ -393,6 +402,9 @@ def spawn(
     network=None,
     public_ip=True,
     extend_group=False,
+    vagrant_cpus=None,
+    vagrant_sync=None,
+    vagrant_provision=None,
 ):
     creds_data = None
     if os.path.exists(CLOUD_CONFIG_FPATH):
@@ -469,13 +481,20 @@ def spawn(
             network=network,
             role=role,
             spawned_cb=print_progress_dot,
+            vagrant_cpus=vagrant_cpus,
+            vagrant_sync=vagrant_sync,
+            vagrant_provision=vagrant_provision,
         )
     except ValueError as e:
         print("\nError: Failed to spawn VMs - " + str(e))
         return 1
     print("DONE")
 
-    if public_ip and (not all(vm.public_ips for vm in vms)):
+    if (
+        provider != Providers.VAGRANT
+        and public_ip
+        and (not all(vm.public_ips for vm in vms))
+    ):
         print("Waiting for VMs to get IP addresses...", end="")
         sys.stdout.flush()  # STDOUT is line-buffered
         while not all(vm.public_ips for vm in vms):
@@ -488,6 +507,16 @@ def spawn(
     else:
         vms_info[group_key] = dump_vms_info(vms)
 
+    if provider == Providers.VAGRANT:
+        vmdir = vms[0].vmdir
+        ssh_config = read_json(SSH_CONFIGS_JSON_FPATH)
+        if not ssh_config:
+            ssh_config = {}
+
+        with open(os.path.join(vmdir, "vagrant-ssh-config"), "r") as f:
+            ssh_config[group_key] = f.read()
+        write_json(SSH_CONFIGS_JSON_FPATH, ssh_config)
+
     write_json(CLOUD_STATE_FPATH, vms_info)
     print("Details about the spawned VMs can be found in %s" % CLOUD_STATE_FPATH)
 
@@ -510,6 +539,36 @@ def _delete_saved_group(vms_info, group_name):
     del vms_info[group_name]
 
 
+def _get_cloud_vms(provider, creds, region, group):
+    if creds is None:
+        raise CFRExitError("Missing/incomplete {} credentials".format(provider.upper()))
+    driver = get_cloud_driver(provider, creds, region)
+
+    assert driver is not None
+
+    nodes = driver.list_nodes()
+    for name, vm_info in group.items():
+        if name == "meta":
+            continue
+        vm_uuid = vm_info["uuid"]
+        vm = CloudVM.get_by_uuid(vm_uuid, nodes=nodes)
+        if vm is not None:
+            yield vm
+        else:
+            print("VM '%s' not found in the clouds" % vm_uuid)
+
+
+def _get_vagrant_vms(group):
+    for name, vm_info in group.items():
+        if name == "meta":
+            continue
+        vm = VagrantVM.get_by_info(name, vm_info)
+        if vm is not None:
+            yield vm
+        else:
+            print("VM '%s' not found locally" % name)
+
+
 def destroy(group_name=None):
     if os.path.exists(CLOUD_CONFIG_FPATH):
         creds_data = read_json(CLOUD_CONFIG_FPATH)
@@ -549,92 +608,54 @@ def destroy(group_name=None):
         raise CFRUserError("No saved VMs found in '{}'".format(CLOUD_STATE_FPATH))
 
     to_destroy = []
+    group_names = None
     if group_name:
         if not group_name.startswith("@"):
             group_name = "@" + group_name
         if group_name not in vms_info:
             print("Group '%s' not found" % group_name)
             return 1
 
+        group_names = [group_name]
+    else:
+        group_names = [key for key in vms_info.keys() if key.startswith("@")]
+
+    ssh_config = read_json(SSH_CONFIGS_JSON_FPATH)
+    assert group_names is not None
+    for group_name in group_names:
         if _is_saved_group(vms_info, group_name):
             _delete_saved_group(vms_info, group_name)
-            write_json(CLOUD_STATE_FPATH, vms_info)
-            return 0
-
-        print("Destroying hosts in the '%s' group" % group_name)
+            continue
 
         region = vms_info[group_name]["meta"]["region"]
         provider = vms_info[group_name]["meta"]["provider"]
-        if provider not in ["aws", "gcp"]:
+        if provider not in ["aws", "gcp", "vagrant"]:
             raise CFRUserError(
-                "Unsupported provider '{}' encountered in '{}', only aws / gcp is supported".format(
+                "Unsupported provider '{}' encountered in '{}', only aws, gcp and vagrant are supported".format(
                     provider, CLOUD_STATE_FPATH
                 )
             )
 
-        driver = None
+        group = vms_info[group_name]
+        vms = []
         if provider == "aws":
-            if aws_creds is None:
-                raise CFRExitError("Missing/incomplete AWS credentials")
-            driver = get_cloud_driver(Providers.AWS, aws_creds, region)
+            vms = _get_cloud_vms(Providers.AWS, aws_creds, region, group)
         if provider == "gcp":
-            if gcp_creds is None:
-                raise CFRExitError("Missing/incomplete GCP credentials")
-            driver = get_cloud_driver(Providers.GCP, gcp_creds, region)
-        assert driver is not None
+            vms = _get_cloud_vms(Providers.GCP, gcp_creds, region, group)
+        if provider == "vagrant":
+            vms = _get_vagrant_vms(group)
 
-        nodes = driver.list_nodes()
-        for name, vm_info in vms_info[group_name].items():
-            if name == "meta":
-                continue
-            vm_uuid = vm_info["uuid"]
-            vm = VM.get_by_uuid(vm_uuid, nodes=nodes)
-            if vm is not None:
-                to_destroy.append(vm)
-            else:
-                print("VM '%s' not found in the clouds" % vm_uuid)
-        del vms_info[group_name]
-    else:
-        print("Destroying all hosts")
-        for group_name in [key for key in vms_info.keys() if key.startswith("@")]:
-            if _is_saved_group(vms_info, group_name):
-                _delete_saved_group(vms_info, group_name)
-                continue
+        for vm in vms:
+            to_destroy.append(vm)
 
-            region = vms_info[group_name]["meta"]["region"]
-            provider = vms_info[group_name]["meta"]["provider"]
-            if provider not in ["aws", "gcp"]:
-                raise CFRUserError(
-                    "Unsupported provider '{}' encountered in '{}', only aws / gcp is supported".format(
-                        provider, CLOUD_STATE_FPATH
-                    )
-                )
+        del vms_info[group_name]
 
-            driver = None
-            if provider == "aws":
-                if aws_creds is None:
-                    raise CFRExitError("Missing/incomplete AWS credentials")
-                driver = get_cloud_driver(Providers.AWS, aws_creds, region)
-            if provider == "gcp":
-                if gcp_creds is None:
-                    raise CFRExitError("Missing/incomplete GCP credentials")
-                driver = get_cloud_driver(Providers.GCP, gcp_creds, region)
-            assert driver is not None
-
-            nodes = driver.list_nodes()
-            for name, vm_info in vms_info[group_name].items():
-                if name == "meta":
-                    continue
-                vm_uuid = vm_info["uuid"]
-                vm = VM.get_by_uuid(vm_uuid, nodes=nodes)
-                if vm is not None:
-                    to_destroy.append(vm)
-                else:
-                    print("VM '%s' not found in the clouds" % vm_uuid)
-            del vms_info[group_name]
+        if ssh_config and group_name in ssh_config:
+            del ssh_config[group_name]
 
     destroy_vms(to_destroy)
     write_json(CLOUD_STATE_FPATH, vms_info)
+    write_json(SSH_CONFIGS_JSON_FPATH, ssh_config)
     return 0
 
 
@@ -655,6 +676,11 @@ def list_platforms():
     return 0
 
 
+def list_boxes():
+    result = subprocess.run(["vagrant", "box", "list"])
+    return result.returncode
+
+
 def init_cloud_config():
     if os.path.exists(CLOUD_CONFIG_FPATH):
         print("File %s already exists" % CLOUD_CONFIG_FPATH)
@@ -1010,7 +1036,7 @@ def connect_cmd(hosts):
         raise CFRExitError("You can only connect to one host at a time")
 
     print("Opening a SSH command shell...")
-    r = subprocess.run(["ssh", hosts[0]])
+    r = subprocess.run(["ssh", "-F", SSH_CONFIG_FPATH, hosts[0]])
     if r.returncode == 0:
         return 0
     if r.returncode < 0: