fifo
bosh

For some time, after watching people having lots of fun with Bosh –  I decided to give “Bosh” the service deployment system a try.

Bosh has great documentation that explains how it can be used on some popular cloud infrastructures. For simplicity sake, lets just call them “clouds”.

That’s where the problems started, because as we all know too well, the officially supported mainstream clouds are Amazon, OpenStack and VMWare. I immediately rejected the idea of replacing my existing and quite functional cloud orchestration system with anything else. I use Project-FiFo which does an absolutely amazing job orchestrating SmartOS hypervisors.

Despite having documentation for specific cloud software,  Bosh is designed not to be tied to any specific cloud or cloud OS. Currently the integration between Bosh and clouds, called CPI (Cloud Provider Interface), lives inside bosh itself which is a problem if you want to add a new one. The good news is that there is a ongoing initiative to support ‘external CPIs’ which aims to decouple Bosh from the cloud implementation itself. The bad news is that its not yet finished and there is no documentation.

This post is about integrating Bosh with FiFo using an external Cloud Provider Interface (CPI).

Each CPI must implement a set of pre-defined inputs and outputs, which is used by Bosh to talk to each cloud system to perform actions: e.g. VM Creation. Each VM is based on a specific OS image, flavour, dataset or “stemcells” as Bosh likes to call them.

1) External CPI

Bosh communicates with external CPI’s using standard input output (stdin, stdout), so you can write external CPI’s in almost any language.

This implementation is called bosh-fifo-cpi and is done using node.js.

It’s used in 2 types of cases:

  1. When deploying “micro bosh” from a client, so that the micro bosh deployer knows how to upload a “stemcell” and create the initial micro bosh instance.
  2. Inside a running “micro bosh” VM, so that it can interact with the Project-FiFo cloud when deploying services.

These are the various methods “Bosh” need from CPI:

  • create_stemcell: Called when uploading a stemcell to the cloud.
  • create_vm
  • delete_stemcell
  • delte_vm
  • get_vm: After bosh creates a vm, it calls this for getting additional VM info e.g. IP address
  • has_vm: Used when triggering bosh cck
  • ping
  • reboot_vm
  • set_vms_metadata: Used to set VM metadata, like the job its running, or its agent id.

Additionally, there are the disk methods such as “create”, “attach”, “detach” and “delete” that are used to mount the data of one VM into another (Bosh migration process) or when you tell Bosh to create a VM with a persistent disk. In the case of FiFo/Smartos we don’t have that functionality so we will just skip them for now.

2) Stemcell

A stemcell is essentially an OS image, which contains some of the base software preinstalled, including the bosh-agent. Bosh need a “stemcell” to initiate the VMs that runs the software its supposed to run, as described in the manifests.

Bosh needs to upload its stemcells to the cloud.

There are currently two ways to accomplish this:

  1. Build a procedure that executes scripts in a chroot, and packs an image from it.
  2. Directly download one from here

All the supported mainstream cloud software have the same thing in common. They manage Linux VM’s. Therefore the existing pre-built stemcells are for Linux. To make Bosh work on another Operating System like SmartOS & Zones will require some modifications to the bosh-agent and a different stemcell build procedure. SmartOS can run KVM perfectly fine, so lets use Linux as a base for now…

We will use the OpenStack Ubuntu “GO” agent image. It contains an OpenStack specific configuration inside. We will change that using a custom init script, which will be executed on every VM first boot, by some additional SmartOS bits we need to put inside.

This process has 3 parts:

  1. Run a VM, using the downloaded OpenStack stemcell
  2. Customize the VM
  3. Build a customized stemcell out of it.

2.1) Running a VM from the original stemcell

We will take the OpenStack stemcell, extract its image and run a VM from it.

In the Global Zone:

 
$ mkdir /opt/extract; cd /opt/extract
$ tar zxf ../bosh-stemcell-you-downloaded.tgz
$ tar zxf image
$ qemu-img info root.img
image: root.img
file format: qcow2
virtual size: 10G (10737418240 bytes)
disk size: 1.0G
cluster_size: 65536
$ qemu-img convert -O raw root.img image.raw
$ qemu-img info image.raw
image: image.raw
file format: raw
virtual size: 10G (10737418240 bytes)
disk size: 1.0G
$ vmadm create -f VM_BOSH.json
Successfully created VM e26d89c6-be16-46ed-a58d-b8d0f7c105ba
$ VM=e26d89c6-be16-46ed-a58d-b8d0f7c105ba
$ DISK=$(vmadm info $VM | json block.0.inserted.file)
$ vmadm stop -f $VM
$ dd if=./image.raw of=$DISK bs=1M
10240+0 records in
10240+0 records out
$ vmadm start $VM
Successfully started VM 664d3956-f8fc-48a1-9f3e-a72ef09ae1a9
$ svcadm restart chunter

This is an example of VM_BOSH.json.

{
        "brand": "kvm",
        "alias": "bosh",
        "resolvers": ["10.0.0.1"],
        "ram": 2048,
        "vcpus": 2,
        "nics": [
          {
                  "nic_tag": "admin",
                  "ip": "10.0.0.240",
                  "netmask": "255.255.255.0",
                  "gateway": "10.0.0.1",
                  "model": "virtio",
                  "primary": true
          }
        ],
        "disks": [
          {
                  "boot": true,
                  "model": "virtio",
                  "size": 10240
          }
        ]
}

2.2) Customise the VM

We will add SmartOS init script into the running VM. The stemcells have the ssh login disabled for root, lets temporary enable it.

Login as root into the machines via FiFo UI or VNC, password is c1oudc0w. Change the last two lines of /etc/ssh/sshd_config to yes, and then service sshd restart

In your notebook:

$ git clone https://github.com/killfill/smartos-ize
$ cd ubuntu; deploy.sh $IP_OF_THE_VM (It will ask you for password more than ones...)

Close the ssh access for root that was opened with VNC above.

2.3) Generating the Stemcell

We should be ready now to transform the created VM running from the OpenStack Bosh stemcell with the above changes, into a stemcell.

  • Power off the VM
  • Snapshot the VM
  • Create an Dataset based on the snapshot. (Imaging tab on the UI of FiFo)
  • Check the Datasets page and confirm that the dataset is available and ready.

Now that the dataset is ready, we can proceed to pack it into a stemcell archive:

  • Dump the dataset into a file, using nfifo:
    • Get the UUID of the created dataset
    • nfifo --json false datasets <uuid> dataset.gz > bosh-dataset.zvol.gz
  • Pack up the Bosh stemcell archive
    • Put the same bosh openstack stemcell we downloaded at the beginning, into a dir
    • mkdir tmp; cd tmp
    • Modify the file apply_spec.json so it contains:
      • {"external_cpi": {"enabled": true, "name": "fifo-cpi"}}
      • This way the file /var/vcap/micro/apply_spec.json in the micro bosh VM gets written that way so bosh-director uses our external CPI instead of the build in OpenStack CPI.
    • $ sed -i.bak "s|openstack|fifo|g" stemcell.MF; rm stemcell.MF.bak
$ cp ../bosh-dataset.zvol.gz ./image
$ tar -zcf ../bosh-stemcell-[version]-fifo-kvm-ubuntu.tgz *

3) Deploy Micro Bosh

This is pretty much following the existing bosh micro deployment documentation

$ gem install bosh_cli_plugin_micro
$ mkdir -p ~/deployments/microbosh-fifo
$ cd ~/deployments/microbosh-fifo

We will need to patch the micro and cpi plugin. Do it as described at the bottom.

Then, create a bosh micro deployment manifest file like this one. Check that the net_id, instance_type and cpi_path matches yours.

Trigger the deployment:

$ git clone https://github.com/killfill/bosh-fifo-cpi.git /path/to/bosh-fifo-cpi
$ bosh micro deployment ~/deployments/microbosh-fifo/micro_bosh.yml
$ bosh micro deploy bosh-stemcell-[version]-fifo-kvm-ubuntu.tgz

That will take the stemcell, upload it to FiFo, create a VM based on it, apply bosh specs, and wait for the components to start.

Logs will be written in ~/deployments/microbosh-fifo/bosh_micro_deploy.log

Now lets put the bosh-fifo-cpi inside just created micro bosh VM, so it can interact with our Cloud. There is a deployment script of bosh-fifo-cpi that does it:

$ ./bin/deploy.sh $ip_of_microbosh

Some of the below checks should now work:

$ bosh micro status
$ bosh micro agent ping
$ bosh target $ip_of_microbosh #user: admin, password: admin
$ bosh status

In case you want to delete the micro bosh vm, bosh micro delete

4) Using the deployed Micro Bosh instance

What makes Bosh interesting is that its not only a configuration management tool like Chef or Puppet, but a service deployment tool. This means that you describe the deployment of the complete service(s) in one manifest file.
Bosh knows exactly how the service is setup, so for example if we have a Redis cluster running and one of the VM dies, Bosh will automatically recreate another one to maintain the desire state of the deployment. The scale of the service is pretty trivial too.

So, lets deploy Redis. First we need to have a stemcell available to use:

$ bosh upload stemcell /path/to/bosh-stemcell-2657-fifo-kvm-ubuntu-go_agent.tgz

Verifying stemcell...
File exists and readable                                     OK
Verifying tarball...
Read tarball                                                 OK
Manifest exists                                              OK
Stemcell image file                                          OK
Stemcell properties                                          OK

Stemcell info
-------------
Name:    bosh-fifo-kvm-ubuntu-go_agent
Version: 2657

Checking if stemcell already exists...
No

Uploading stemcell...

bosh-stemcell:  20% |ooooooooooooooooooooooooooooooooooooooo| 549.4MB   2.6MB/s Time: 00:03:31

Director task 2
  Started update stemcell
  Started update stemcell &gt; Extracting stemcell archive. Done (00:00:11)
  Started update stemcell &gt; Verifying stemcell manifest. Done (00:00:00)
  Started update stemcell &gt; Checking if this stemcell already exists. Done (00:00:00)
  Started update stemcell &gt; Uploading stemcell bosh-fifo-kvm-ubuntu-go_agent/2657 to the cloud. Done (00:00:00)
  Started update stemcell &gt; Save stemcell bosh-fifo-kvm-ubuntu-go_agent/2657 (8f38dca2-9497-4807-a5f1-62aa1d6a7728). Done (00:00:00)
     Done update stemcell (00:00:11)

Task 2 done

Started        2014-08-11 01:19:03 UTC
Finished    2014-08-11 01:19:14 UTC
Duration    00:00:11

Stemcell uploaded and created.

bosh stemcells should list the just uploaded stemcell.

Create a redis deployment manifest like this one. It can be simplified but just let use it:

compilation:
  cloud_properties:
    instance_type: 8ade6653-86ac-4777-82a3-728ac5751142
  network: redis1
  reuse_compilation_vms: true
  workers: 6
director_uuid: 51c09760-514c-4353-8a8a-4e8c3b4019fb
jobs:
- instances: 1
  name: redis_leader_z1
  networks:
  - name: redis1
    static_ips: null
  persistent_disk: 0
  resource_pool: small_z1
  templates:
  - name: redis
    release: redis
- instances: 2
  name: redis_z1
  networks:
  - name: redis1
    static_ips: null
  persistent_disk: 0
  properties:
    redis:
      master: 0.redis-leader-z1.redis1.a-redis-cluster.microbosh
  resource_pool: small_z1
  templates:
  - name: redis
    release: redis
  update:
    canaries: 10
meta:
  environment: a-redis-cluster-huh
  stemcell:
    name: bosh-fifo-kvm-ubuntu-go_agent
    version: latest
name: a-redis-cluster
networks:
- cloud_properties: {}
  name: floating
  type: vip
- cloud_properties:
    net_id: c0c15666-0b72-4998-aaf2-2bb40e71b799
    security_groups:
    - default
  name: redis1
  type: dynamic
properties:
  consul: null
releases:
- name: redis
  version: latest
resource_pools:
- cloud_properties:
    instance_type: 0a827ade-f24c-4cb3-9df4-3ad35561fa9c
  name: small_z1
  network: redis1
  size: 3
  stemcell:
    name: bosh-fifo-kvm-ubuntu-go_agent
    version: latest
update:
  canaries: 1
  canary_watch_time: 1000-30000
  max_in_flight: 50
  update_watch_time: 1000-30000

Then run:

bosh deployment redis-fifo-manifest.yml
bosh upload release https://redis-boshrelease.s3.amazonaws.com/boshrelease-redis-5.tgz
bosh -n deploy

Done!

Here is a video that goes throw the process of the Redis deployment and then scale it.

Lets try something a little more complex. Wipe out the Redis cluster just deployed, create a Cloud Foundry manifest like this one. Again, this is not the simplest manifest, but it can be simplified.

compilation:
  cloud_properties:
    instance_type: 716a4b06-7031-4066-aab0-31758baf2808
  network: cf1
  reuse_compilation_vms: true
  workers: 6
director_uuid: d1581edd-b03b-4486-b417-0d24a7e4cd25
jobs:
- instances: 1
  name: nats_z1
  networks:
  - name: cf1
    static_ips: null
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: nats
    release: cf
  - name: nats_stream_forwarder
    release: cf
- instances: 0
  name: nats_z2
  networks:
  - name: cf2
    static_ips: []
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: nats
    release: cf
  - name: nats_stream_forwarder
    release: cf
- instances: 1
  name: etcd_z1
  networks:
  - name: cf1
    static_ips: null
  persistent_disk: 0
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: etcd
    release: cf
  - name: etcd_metrics_server
    release: cf
- instances: 0
  name: etcd_z2
  networks:
  - name: cf2
    static_ips: []
  persistent_disk: 0
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: etcd
    release: cf
  - name: etcd_metrics_server
    release: cf
- instances: 0
  name: logs_z1
  networks:
  - name: cf1
    static_ips: null
  persistent_disk: 0
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: syslog_aggregator
    release: cf
- instances: 0
  name: logs_z2
  networks:
  - name: cf2
    static_ips: []
  persistent_disk: 0
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: syslog_aggregator
    release: cf
- instances: 0
  name: stats_z1
  networks:
  - name: cf1
  properties:
    networks:
      apps: cf1
  resource_pool: small_z1
  templates:
  - name: collector
    release: cf
- instances: 1
  name: nfs_z1
  networks:
  - name: cf1
    static_ips: null
  persistent_disk: 0
  resource_pool: medium_z1
  templates:
  - name: debian_nfs_server
    release: cf
- instances: 1
  name: postgres_z1
  networks:
  - name: cf1
    static_ips: null
  persistent_disk: 0
  resource_pool: medium_z1
  templates:
  - name: postgres
    release: cf
- instances: 1
  name: uaa_z1
  networks:
  - name: cf1
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: uaa
    release: cf
- instances: 0
  name: uaa_z2
  networks:
  - name: cf2
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: uaa
    release: cf
- instances: 1
  name: login_z1
  networks:
  - name: cf1
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: login
    release: cf
- instances: 0
  name: login_z2
  networks:
  - name: cf2
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: login
    release: cf
- instances: 1
  name: api_z1
  networks:
  - name: cf1
  persistent_disk: 0
  properties:
    metron_agent:
      zone: z1
    networks:
      apps: cf1
    nfs_server:
      address: 0.nfs-z1.cf1.cf-baby.microbosh
      allow_from_entries:
      - 10.0.0.0/24
      - null
      - null
      share: null
  resource_pool: large_z1
  templates:
  - name: cloud_controller_ng
    release: cf
  - name: metron_agent
    release: cf
- instances: 0
  name: api_z2
  networks:
  - name: cf2
  persistent_disk: 0
  properties:
    metron_agent:
      zone: z2
    networks:
      apps: cf2
    nfs_server:
      address: 0.nfs-z1.cf1.cf-baby.microbosh
      allow_from_entries:
      - 10.0.0.0/24
      - null
      - null
      share: null
  resource_pool: large_z2
  templates:
  - name: cloud_controller_ng
    release: cf
  - name: metron_agent
    release: cf
- instances: 1
  name: clock_global
  networks:
  - name: cf1
  persistent_disk: 0
  properties:
    metron_agent:
      zone: z1
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: cloud_controller_clock
    release: cf
  - name: metron_agent
    release: cf
- instances: 1
  name: api_worker_z1
  networks:
  - name: cf1
  persistent_disk: 0
  properties:
    metron_agent:
      zone: z1
    networks:
      apps: cf1
    nfs_server:
      address: 0.nfs-z1.cf1.cf-baby.microbosh
      allow_from_entries:
      - 10.0.0.0/24
      - null
      - null
      share: null
  resource_pool: small_z1
  templates:
  - name: cloud_controller_worker
    release: cf
  - name: metron_agent
    release: cf
- instances: 0
  name: api_worker_z2
  networks:
  - name: cf2
  persistent_disk: 0
  properties:
    metron_agent:
      zone: z2
    networks:
      apps: cf2
    nfs_server:
      address: 0.nfs-z1.cf1.cf-baby.microbosh
      allow_from_entries:
      - 10.0.0.0/24
      - null
      - null
      share: null
  resource_pool: small_z2
  templates:
  - name: cloud_controller_worker
    release: cf
  - name: metron_agent
    release: cf
- instances: 1
  name: hm9000_z1
  networks:
  - name: cf1
  properties:
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: hm9000
    release: cf
- instances: 0
  name: hm9000_z2
  networks:
  - name: cf2
  properties:
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: hm9000
    release: cf
- instances: 1
  name: runner_z1
  networks:
  - name: cf1
    static_ips: null
  properties:
    metron_agent:
      zone: z1
    networks:
      apps: cf1
  resource_pool: runner_z1
  templates:
  - name: dea_next
    release: cf
  - name: dea_logging_agent
    release: cf
  - name: metron_agent
    release: cf
  update:
    max_in_flight: 1
- instances: 0
  name: runner_z2
  networks:
  - name: cf2
    static_ips: null
  properties:
    metron_agent:
      zone: z2
    networks:
      apps: cf2
  resource_pool: runner_z2
  templates:
  - name: dea_next
    release: cf
  - name: dea_logging_agent
    release: cf
  - name: metron_agent
    release: cf
  update:
    max_in_flight: 1
- instances: 1
  name: loggregator_z1
  networks:
  - name: cf1
    static_ips: null
  properties:
    loggregator:
      zone: z1
    networks:
      apps: cf1
  resource_pool: medium_z1
  templates:
  - name: loggregator
    release: cf
- instances: 0
  name: loggregator_z2
  networks:
  - name: cf2
    static_ips: []
  properties:
    loggregator:
      zone: z2
    networks:
      apps: cf2
  resource_pool: medium_z2
  templates:
  - name: loggregator
    release: cf
- instances: 1
  name: loggregator_trafficcontroller_z1
  networks:
  - name: cf1
    static_ips: null
  properties:
    networks:
      apps: cf1
    traffic_controller:
      zone: z1
  resource_pool: small_z1
  templates:
  - name: loggregator_trafficcontroller
    release: cf
- instances: 0
  name: loggregator_trafficcontroller_z2
  networks:
  - name: cf2
    static_ips: []
  properties:
    networks:
      apps: cf2
    traffic_controller:
      zone: z2
  resource_pool: small_z2
  templates:
  - name: loggregator_trafficcontroller
    release: cf
- instances: 1
  name: router_z1
  networks:
  - name: cf1
    static_ips: null
  properties:
    metron_agent:
      zone: z1
    networks:
      apps: cf1
  resource_pool: router_z1
  templates:
  - name: gorouter
    release: cf
  - name: metron_agent
    release: cf
- instances: 0
  name: router_z2
  networks:
  - name: cf2
    static_ips: []
  properties:
    metron_agent:
      zone: z2
    networks:
      apps: cf2
  resource_pool: router_z2
  templates:
  - name: gorouter
    release: cf
  - name: metron_agent
    release: cf
- default_networks:
  - name: cf1
    static_ips: null
  instances: 1
  name: ha_proxy_z1
  networks:
  - default:
    - dns
    - gateway
    name: cf1
    static_ips: null
  properties:
    ha_proxy:
      ssl_pem: '-----BEGIN CERTIFICATE-----

        MIIDYTCCAkmgAwIBAgIJAOQy8l+wDKK9MA0GCSqGSIb3DQEBBQUAMCkxDTALBgNV

        BAoTBEJvc2gxGDAWBgNVBAMUDyouaG9sYS52aXJ0dS5jbDAeFw0xNDA4MjAwMTMx

        NTdaFw0yMDAyMTAwMTMxNTdaMCkxDTALBgNVBAoTBEJvc2gxGDAWBgNVBAMUDyou

        aG9sYS52aXJ0dS5jbDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAMzt

        DfGw/sM6oF4hBkLnp0yXaq9XbnC8lt4VpRUatnt+TXsV6Dc2bEQN9qsf/xs2Ocpc

        6eRHT2PLkpNKPVg9q3FDLUhaHT7eoHihgAJzpqXvz9jkra3CvdDrabPKi+Ym5iiG

        BJPJ0hOs8jGRFIEFiEob3kk07sAN9ekh8LJJ5uZrveJfnn9CGyVr2h907AwEESuk

        jWw/gY1n4RBIMUDhv66FIK2N/hxcgwdyQJKpRJMpsEB+Y/rPMZ3mN8/hY2pJn+Qv

        dXF7RNzDYpNbbJ+n22e+whbv38ETMGnfd2KMMX+MisKFYSLGMWxBnmXaEqc+JVhK

        DNd/ozOfnRCAv8nsNEkCAwEAAaOBizCBiDAdBgNVHQ4EFgQUxsxwoXBIYOscHlJU

        Gl0g1o/vo+kwWQYDVR0jBFIwUIAUxsxwoXBIYOscHlJUGl0g1o/vo+mhLaQrMCkx

        DTALBgNVBAoTBEJvc2gxGDAWBgNVBAMUDyouaG9sYS52aXJ0dS5jbIIJAOQy8l+w

        DKK9MAwGA1UdEwQFMAMBAf8wDQYJKoZIhvcNAQEFBQADggEBAD3sOQa+ux5qGhfx

        TbWmDMGYITO7W/4PSiaoOP9dDqOeAREy8I+S0TWKZD8eibiCSTIEIrEzd+fVaQzd

        +QYCQ/7fIGx+UProqN0I0CD8tNa268AwhvD2HFApZ3D1L45WKZJwjovrH36a9TWz

        Eqh3zwAjVBVAieu6FcDRK6AW4pY/P2CYI2dst4WJ6UI6LDVWbQgjl7K6j4tIqaqH

        RFZd3ws8w/LivBALDYwiLI7RZ60cZf5FrsD+3wz8fzXqzgdd1GPFGS0R3E0evn6k

        /deo/FfZ7p5PbBT6gi2tFMnM7yelfhDWhtuEPaQHzg4ZvXMQ8uXdM7h1wETJtvkN

        Zy1oiEQ=

        -----END CERTIFICATE-----

        -----BEGIN RSA PRIVATE KEY-----

        MIIEogIBAAKCAQEAzO0N8bD+wzqgXiEGQuenTJdqr1ducLyW3hWlFRq2e35NexXo

        NzZsRA32qx//GzY5ylzp5EdPY8uSk0o9WD2rcUMtSFodPt6geKGAAnOmpe/P2OSt

        rcK90Otps8qL5ibmKIYEk8nSE6zyMZEUgQWIShveSTTuwA316SHwsknm5mu94l+e

        f0IbJWvaH3TsDAQRK6SNbD+BjWfhEEgxQOG/roUgrY3+HFyDB3JAkqlEkymwQH5j

        +s8xneY3z+Fjakmf5C91cXtE3MNik1tsn6fbZ77CFu/fwRMwad93Yowxf4yKwoVh

        IsYxbEGeZdoSpz4lWEoM13+jM5+dEIC/yew0SQIDAQABAoIBAA3KHdgXqu0EMupi

        moNCB1X2Em2+GmZnqldDaZpj88eXliskLD4a4Woq7mBRhN8V5Qmi+vIrgkIa3sAe

        jMY1UCYM+S0XM0SZwn6bj3uawKoF2xyjNrSsmA1I6hH5HTTdhTtQ2VoAiKxeROXk

        91Y/Qw6JlYmjkDVIOzX23LWCl5EFrCAMzu7g3u0penvKF8FPFiwdX7W2XuovZhdF

        5cFDBZElrjZEPLKUaFp+glUhRh7O2ywhCCSziqI8YDIvB9JidmQV7K9GI/z5jAyM

        xQXQd2f7UqzKd5YhuQ9FGmKxHdgv/5AMjVVC3m8s1qbrr+zLCK2u0X1E6XfyaxOV

        kus5hAECgYEA8g9OrxEAJ/DlC1meyyNspkGhsC74b5+6+0N2gj+j0EpirdMtt3MC

        LA5SIC0LWxx8e/513X7/dcxN2q2dgRS7f9iTdRCrF/F0obJLuJrRa7XVhL4Qs6KD

        0gXZx3fTPDm1RM0qcx0x+E6Veh9dPV5R6TaMrnt3vg+5yRxZtZurXQkCgYEA2LpI

        X7ek4NwI+eNojoqeLnj004o/QUPNnywjPfj0KeVH/G8wnoBr7seANYxNCLCNvgLc

        +Nt+94HGqZxycJcikl8u0q5vn4+dpFAfDZ0wsOrB9bT9Gqtt3OG/svKHs15Y2ulj

        58Nj8XrToCrzyUx3WmRpAxmZNYwpdGdJJNZXLUECgYAZwxVNvpJ0rcfCyHwFBLUo

        kK2jsawRPT/KRCjVve4yWehctbAU6TGe/Lsy33ho7wQ0siVbOIv5vo++qWGs/kbi

        WWuANxDukG0oVHCIqjmFrCmYYh4m8DH1ZVgDSbZ6n22yb9+HtuC9m6d/At2KjYJM

        PtkELPHWkZflVpYZpQK8+QKBgGd3x8WqrxSK994VK32hwBL/vCKiBcDq1ef/8nNL

        v4Wg6NUjYFrLhLviktr9uGlFoOXuCADnCZ6wXl/rSNdn0CGBP6FllfTfY6tmdOzS

        9sP/z8m2Nn/dB/5D6HlOQGrpyBH7Cq9DIRvQP+hfmTOpRfHprhjEx9hAqinghZMI

        oWYBAoGAShB1ebq0IlIHizbHNZfWSt143cSYLaxS/kxhMp5gLQRnokAeXZ8jMlOj

        tkr6YK8KXI4ueTfvNckHFqGeny134zo487CP/Z68o2mTJTvzbLB8etV1mXuQ3QJN

        4knZIM/OXK56ae5ydydaYcAWTo3LV8VvqJIhyoWFryLM4oSSud4=

        -----END RSA PRIVATE KEY-----

'
    networks:
      apps: cf1
    router:
      servers:
        z1:
        - 0.router-z1.cf1.cf-baby.microbosh
        z2: []
  resource_pool: router_z1
  templates:
  - name: haproxy
    release: cf
- instances: 0
  lifecycle: errand
  name: acceptance_tests
  networks:
  - name: cf1
  resource_pool: small_errand
  templates:
  - name: acceptance-tests
    release: cf
- instances: 0
  lifecycle: errand
  name: smoke_tests
  networks:
  - name: cf1
  resource_pool: small_errand
  templates:
  - name: smoke-tests
    release: cf
meta:
  environment: cf-baby
  releases:
  - name: cf
    version: latest
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
name: cf-baby
networks:
- cloud_properties:
    net_id: c0c15666-0b72-4998-aaf2-2bb40e71b799
  name: cf1
  type: dynamic
- cloud_properties:
    net_id: c0c15666-0b72-4998-aaf2-2bb40e71b799
  name: cf2
  type: dynamic
properties:
  acceptance_tests: null
  app_domains:
  - hola.virtu.cl
  cc:
    app_events:
      cutoff_age_in_days: 31
    app_usage_events:
      cutoff_age_in_days: 31
    audit_events:
      cutoff_age_in_days: 31
    billing_event_writing_enabled: true
    broker_client_timeout_seconds: 70
    buildpacks:
      buildpack_directory_key: hola.virtu.cl-cc-buildpacks
      cdn: null
      fog_connection: null
    bulk_api_password: password
    client_max_body_size: 1536M
    db_encryption_key: password
    default_app_disk_in_mb: 1024
    default_app_memory: 1024
    default_buildpacks:
    - name: java_buildpack
      package: buildpack_java
    - name: ruby_buildpack
      package: buildpack_ruby
    - name: nodejs_buildpack
      package: buildpack_nodejs
    - name: go_buildpack
      package: buildpack_go
    - name: python_buildpack
      package: buildpack_python
    - name: php_buildpack
      package: buildpack_php
    default_quota_definition: default
    default_running_security_groups:
    - public_networks
    - dns
    default_staging_security_groups:
    - public_networks
    - dns
    development_mode: false
    diego: false
    directories: null
    disable_custom_buildpacks: false
    droplets:
      cdn: null
      droplet_directory_key: hola.virtu.cl-cc-droplets
      fog_connection: null
    external_host: api
    install_buildpacks:
    - name: java_buildpack
      package: buildpack_java
    - name: ruby_buildpack
      package: buildpack_ruby
    - name: nodejs_buildpack
      package: buildpack_nodejs
    - name: go_buildpack
      package: buildpack_go
    - name: python_buildpack
      package: buildpack_python
    - name: php_buildpack
      package: buildpack_php
    jobs:
      app_bits_packer:
        timeout_in_seconds: null
      app_events_cleanup:
        timeout_in_seconds: null
      app_usage_events_cleanup:
        timeout_in_seconds: null
      blobstore_delete:
        timeout_in_seconds: null
      blobstore_upload:
        timeout_in_seconds: null
      droplet_deletion:
        timeout_in_seconds: null
      droplet_upload:
        timeout_in_seconds: null
      global:
        timeout_in_seconds: 14400
      model_deletion:
        timeout_in_seconds: null
    maximum_app_disk_in_mb: 2048
    newrelic:
      capture_params: false
      developer_mode: false
      environment_name: cf-baby
      license_key: null
      monitor_mode: false
      transaction_tracer:
        enabled: true
        record_sql: obfuscated
    packages:
      app_package_directory_key: hola.virtu.cl-cc-packages
      cdn: null
      fog_connection: null
      max_package_size: 1073741824
    quota_definitions:
      default:
        memory_limit: 10240
        non_basic_services_allowed: true
        total_routes: 1000
        total_services: 100
    resource_pool:
      cdn: null
      fog_connection: null
      resource_directory_key: hola.virtu.cl-cc-resources
    security_group_definitions:
    - name: public_networks
      rules:
      - destination: 0.0.0.0-9.255.255.255
        protocol: all
      - destination: 11.0.0.0-169.253.255.255
        protocol: all
      - destination: 169.255.0.0-172.15.255.255
        protocol: all
      - destination: 172.32.0.0-192.167.255.255
        protocol: all
      - destination: 192.169.0.0-255.255.255.255
        protocol: all
    - name: dns
      rules:
      - destination: 0.0.0.0/0
        ports: "53"
        protocol: tcp
      - destination: 0.0.0.0/0
        ports: "53"
        protocol: udp
    srv_api_uri: https://api.hola.virtu.cl
    stacks: null
    staging_upload_password: upload-password
    staging_upload_user: upload-user
    system_buildpacks:
    - name: java_buildpack
      package: buildpack_java
    - name: ruby_buildpack
      package: buildpack_ruby
    - name: nodejs_buildpack
      package: buildpack_nodejs
    - name: go_buildpack
      package: buildpack_go
    - name: python_buildpack
      package: buildpack_python
    - name: php_buildpack
      package: buildpack_php
    user_buildpacks: []
  ccdb:
    address: 0.postgres-z1.cf1.cf-baby.microbosh
    databases:
    - name: ccdb
      tag: cc
    db_scheme: postgres
    port: 5524
    roles:
    - name: ccadmin
      password: ccadmin
      tag: admin
  collector: null
  databases:
    address: 0.postgres-z1.cf1.cf-baby.microbosh
    databases:
    - citext: true
      name: ccdb
      tag: cc
    - citext: true
      name: uaadb
      tag: uaa
    db_scheme: postgres
    port: 5524
    roles:
    - name: ccadmin
      password: ccadmin
      tag: admin
    - name: uaaadmin
      password: uaaadmin
      tag: admin
  dea_next:
    allow_networks: null
    deny_networks: null
    directory_server_protocol: https
    disk_mb: 10000
    disk_overcommit_factor: 2
    evacuation_bail_out_time_in_seconds: 600
    instance_disk_inode_limit: 200000
    kernel_network_tuning_enabled: true
    logging_level: debug
    memory_mb: 1024
    memory_overcommit_factor: 3
    staging_disk_inode_limit: 200000
    staging_disk_limit_mb: 4096
    staging_memory_limit_mb: 1024
  description: La Nube
  disk_quota_enabled: true
  domain: hola.virtu.cl
  dropsonde:
    enabled: true
  etcd:
    machines:
    - 0.etcd-z1.cf1.cf-baby.microbosh
  etcd_metrics_server:
    nats:
      machines:
      - 0.nats-z1.cf1.cf-baby.microbosh
      password: nats
      username: nats
  logger_endpoint:
    port: 4443
  loggregator:
    blacklisted_syslog_ranges: []
    debug: false
    maxRetainedLogMessages: 100
  loggregator_endpoint:
    shared_secret: password
  login:
    analytics:
      code: null
      domain: null
    asset_base_url: null
    brand: oss
    catalina_opts: -Xmx768m -XX:MaxPermSize=256m
    links:
      home: https://console.hola.virtu.cl
      network: null
      passwd: https://console.hola.virtu.cl/password_resets/new
      signup: https://console.hola.virtu.cl/register
      signup-network: null
    protocol: https
    smtp:
      host: null
      password: null
      port: null
      user: null
    tiles: null
    uaa_base: null
    uaa_certificate: '-----BEGIN CERTIFICATE-----

      MIIDYTCCAkmgAwIBAgIJAOQy8l+wDKK9MA0GCSqGSIb3DQEBBQUAMCkxDTALBgNV

      BAoTBEJvc2gxGDAWBgNVBAMUDyouaG9sYS52aXJ0dS5jbDAeFw0xNDA4MjAwMTMx

      NTdaFw0yMDAyMTAwMTMxNTdaMCkxDTALBgNVBAoTBEJvc2gxGDAWBgNVBAMUDyou

      aG9sYS52aXJ0dS5jbDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAMzt

      DfGw/sM6oF4hBkLnp0yXaq9XbnC8lt4VpRUatnt+TXsV6Dc2bEQN9qsf/xs2Ocpc

      6eRHT2PLkpNKPVg9q3FDLUhaHT7eoHihgAJzpqXvz9jkra3CvdDrabPKi+Ym5iiG

      BJPJ0hOs8jGRFIEFiEob3kk07sAN9ekh8LJJ5uZrveJfnn9CGyVr2h907AwEESuk

      jWw/gY1n4RBIMUDhv66FIK2N/hxcgwdyQJKpRJMpsEB+Y/rPMZ3mN8/hY2pJn+Qv

      dXF7RNzDYpNbbJ+n22e+whbv38ETMGnfd2KMMX+MisKFYSLGMWxBnmXaEqc+JVhK

      DNd/ozOfnRCAv8nsNEkCAwEAAaOBizCBiDAdBgNVHQ4EFgQUxsxwoXBIYOscHlJU

      Gl0g1o/vo+kwWQYDVR0jBFIwUIAUxsxwoXBIYOscHlJUGl0g1o/vo+mhLaQrMCkx

      DTALBgNVBAoTBEJvc2gxGDAWBgNVBAMUDyouaG9sYS52aXJ0dS5jbIIJAOQy8l+w

      DKK9MAwGA1UdEwQFMAMBAf8wDQYJKoZIhvcNAQEFBQADggEBAD3sOQa+ux5qGhfx

      TbWmDMGYITO7W/4PSiaoOP9dDqOeAREy8I+S0TWKZD8eibiCSTIEIrEzd+fVaQzd

      +QYCQ/7fIGx+UProqN0I0CD8tNa268AwhvD2HFApZ3D1L45WKZJwjovrH36a9TWz

      Eqh3zwAjVBVAieu6FcDRK6AW4pY/P2CYI2dst4WJ6UI6LDVWbQgjl7K6j4tIqaqH

      RFZd3ws8w/LivBALDYwiLI7RZ60cZf5FrsD+3wz8fzXqzgdd1GPFGS0R3E0evn6k

      /deo/FfZ7p5PbBT6gi2tFMnM7yelfhDWhtuEPaQHzg4ZvXMQ8uXdM7h1wETJtvkN

      Zy1oiEQ=

      -----END CERTIFICATE-----

'
    url: null
  metron_endpoint:
    shared_secret: password
  nats:
    address: 0.nats-z1.cf1.cf-baby.microbosh
    debug: false
    machines:
    - 0.nats-z1.cf1.cf-baby.microbosh
    monitor_port: 0
    password: nats
    port: 4222
    prof_port: 0
    trace: false
    user: nats
  nfs_server:
    address: 0.nfs-z1.cf1.cf-baby.microbosh
    allow_from_entries:
    - 10.0.0.0/24
    - null
    - null
    share: null
  request_timeout_in_seconds: 300
  router:
    status:
      password: router
      user: router
  smoke_tests: null
  ssl:
    skip_cert_verify: true
  support_address: http://www.virtualizado.cl
  syslog_aggregator: null
  system_domain: hola.virtu.cl
  system_domain_organization: null
  uaa:
    admin:
      client_secret: password
    authentication:
      policy:
        countFailuresWithinSeconds: null
        lockoutAfterFailures: null
        lockoutPeriodSeconds: null
    batch:
      password: password
      username: batch
    catalina_opts: -Xmx768m -XX:MaxPermSize=256m
    cc:
      client_secret: password
    clients:
      app-direct:
        access-token-validity: 1209600
        authorities: app_direct_invoice.write
        authorized-grant-types: authorization_code,client_credentials,password,refresh_token,implicit
        override: true
        redirect-uri: https://console.hola.virtu.cl
        refresh-token-validity: 1209600
        secret: password
      developer_console:
        access-token-validity: 1209600
        authorities: scim.write,scim.read,cloud_controller.read,cloud_controller.write,password.write,uaa.admin,uaa.resource,cloud_controller.admin,billing.admin
        authorized-grant-types: authorization_code,client_credentials
        override: true
        redirect-uri: https://console.hola.virtu.cl/oauth/callback
        refresh-token-validity: 1209600
        scope: openid,cloud_controller.read,cloud_controller.write,password.write,console.admin,console.support
        secret: password
      login:
        authorities: oauth.login
        authorized-grant-types: authorization_code,client_credentials,refresh_token
        override: true
        redirect-uri: https://login.hola.virtu.cl
        scope: openid,oauth.approvals
        secret: password
      notifications:
        secret: password
      servicesmgmt:
        secret: password
      space-mail:
        secret: password
      support-services:
        secret: password
    jwt:
      signing_key: '-----BEGIN RSA PRIVATE KEY-----

        MIIEpgIBAAKCAQEAz4OguciBpCHUcYvK0BPYIyVyWyUG2jXG2QhHG0kV3pxInsei

        joVelvwJnmBIxSTwgAoOiLH/ECoyRsNWL8eDfldIwOHmC/HOeQ16nCAXOcm4Oan+

        bz6loZ4szLNVnTucUlAFR8nLTUQDLtJvzcCdh0dEJOJAVJqucUfeKm/LxH+TfW0J

        WlqY2ezwlW7P3wQtD4gAAkWHl0IUdAp5Uv+GbeBiQ80n9DwE821Ih5exKsjZWyaC

        wQAaSMJ/EZhLnF1bH9qXnxz+Y+CL259MRE/gbAGkGTcwUvSMVAJW5P+sdmoACerK

        6nxbtmSv3IUtzOs7nQex1Wh0xNFyyiBAWYdRFQIDAQABAoIBAQCPPxXcKYUUgCXe

        m9lM+eXhnR/ruwuX5jd16jjrpO/r04Zt1PFp4LcDEsixA1Y4gN9kA40032prBjXe

        k0kTgm5K8IXQYCAlbq5Q8fUJgT2HLrB1Xdu9f0lm34GdbM5U06l9SK/cdnhYxfJM

        iAeVyLj1gUSREX3G6WAx4qKCONo4sXl5iYMMm8q/sf5mYu3T249gshZCHscyxhZ2

        Ckhvq8bIEeCkZ35oaFRCT6ZX9FblsMKCDQRlbDvxpKTvf6dJmHpFaIkKGV6HlV5C

        lBNXdZ+xi6ZRemDDebbRDo0GZRk5Emuzn3UNlfAuqOd3KWtFSevS4l2XSX2gEIQB

        eWvE99MBAoGBAPZ1oiJ4HgfUKwJKOuDahi9ZL+ah8uoZSLl/yaPYmiYT7DdAFGok

        V3aEr35I/N32W8K2MmTnGOis8bAWFUoWOYAhvSsM7sQZazY/41AltLqQ14OKmuBt

        OBdVQpGUR2WalP7c3eQtWkLpdBFGzBrklTppALmY6/pg7Buscv7kcXFdAoGBANeM

        Db0/C8T4610Fb026g37jSePvKlFW94makeAPZmGNSS4xcXiYZjpC5fDFcq1ko+iu

        ApyXq4VHbf4s3/8ps2I5oPD4S5qqIukY5Elf+FFR17y1W4S9QETyxH1ak5xK/mY6

        nQlAQzOek99dicxuFBX0bv8nNvYaep3cdS6iYUsZAoGBAJl6b9lfS/Q8H/xxFXFp

        JB7gP4A0Yf1+eVdrSLqL7JLB3Vwav5vJvd3EizTtZusiu8nphEBp2Y0sBeZOGjvk

        0zID0LW6fAYKfkVuS7RhhsaX5F26NvCDk2aUZbwaxstzFE4SZOvvgkIg93K6z+f/

        aR7SvOzXWjy1ltM09PfecvbRAoGBAMjGxo1/SoFBtEtDorfugkoRgWR2BR8bORzO

        BA+HiMBZkM8ccyv4GVEvu2wZH4bv9dXpMLmYdcaCEKzwM/5EY7mdIJMVVadNz5L9

        ArhuUqm86I35CLSObG1NDUYOtuyg1nN3KYvSdkaA8hDf2lm4N7FzrGizqHh5lbmx

        /t1fl+0RAoGBAMaU4lfXY1d6pzLvUUU+lb6l5DOsX9vUCuzZgARmkY7LvK88wZMF

        laTUXazGZAmlRUeKftqK9SEZ/e5s53B5JSdE3rVKMLH6sPxkybbUzIdyeU4RMMcS

        GNvX8WEVeFFRbS3x3QffJhoCQyWK+DBBuNPfxC3Kaq/CUAqIswcq3Vr/

        -----END RSA PRIVATE KEY-----

'
      verification_key: '-----BEGIN PUBLIC KEY-----

        MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAz4OguciBpCHUcYvK0BPY

        IyVyWyUG2jXG2QhHG0kV3pxInseijoVelvwJnmBIxSTwgAoOiLH/ECoyRsNWL8eD

        fldIwOHmC/HOeQ16nCAXOcm4Oan+bz6loZ4szLNVnTucUlAFR8nLTUQDLtJvzcCd

        h0dEJOJAVJqucUfeKm/LxH+TfW0JWlqY2ezwlW7P3wQtD4gAAkWHl0IUdAp5Uv+G

        beBiQ80n9DwE821Ih5exKsjZWyaCwQAaSMJ/EZhLnF1bH9qXnxz+Y+CL259MRE/g

        bAGkGTcwUvSMVAJW5P+sdmoACerK6nxbtmSv3IUtzOs7nQex1Wh0xNFyyiBAWYdR

        FQIDAQAB

        -----END PUBLIC KEY-----

'
    ldap: null
    login:
      client_secret: null
    no_ssl: false
    scim:
      external_groups: null
      userids_enabled: false
      users:
      - admin|admin|scim.write,scim.read,openid,cloud_controller.admin,uaa.admin,password.write
      - services|password|scim.write,scim.read,openid,cloud_controller.admin
    spring_profiles: null
    url: https://uaa.hola.virtu.cl
  uaadb:
    address: 0.postgres-z1.cf1.cf-baby.microbosh
    databases:
    - name: uaadb
      tag: uaa
    db_scheme: postgresql
    port: 5524
    roles:
    - name: uaaadmin
      password: uaaadmin
      tag: admin
releases:
- name: cf
  version: latest
resource_pools:
- cloud_properties:
    instance_type: 0a827ade-f24c-4cb3-9df4-3ad35561fa9c
  name: small_z1
  network: cf1
  size: 2
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: m1.small
  name: small_z2
  network: cf2
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: 0a827ade-f24c-4cb3-9df4-3ad35561fa9c
  name: medium_z1
  network: cf1
  size: 9
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: m1.medium
  name: medium_z2
  network: cf2
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: 8ade6653-86ac-4777-82a3-728ac5751142
  name: large_z1
  network: cf1
  size: 1
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: m1.large
  name: large_z2
  network: cf2
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: 8ade6653-86ac-4777-82a3-728ac5751142
  name: runner_z1
  network: cf1
  size: 1
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: m1.large
  name: runner_z2
  network: cf2
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: 0a827ade-f24c-4cb3-9df4-3ad35561fa9c
  name: router_z1
  network: cf1
  size: 2
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: m1.medium
  name: router_z2
  network: cf2
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
- cloud_properties:
    instance_type: 0a827ade-f24c-4cb3-9df4-3ad35561fa9c
  name: small_errand
  network: cf1
  size: 0
  stemcell:
    name: bosh-fifo-kvm-ubuntu-trusty-go_agent
    version: latest
update:
  canaries: 0
  canary_watch_time: 30000-600000
  max_in_flight: 1
  serial: true
  update_watch_time: 5000-600000

Then deploy Cloud Foundry on FiFo/SmartOS:

$ bosh delete deployment a-redis-cluster
$ bosh upload release https://community-shared-boshreleases.s3.amazonaws.com/boshrelease-cf-178.tgz
$ bosh deployment cf-baby-178.yml
$ bosh -n deploy

bosh-fifo-success

The output can be seen in this boring video

Note: After uploading the video i realize that is has no FiFo/SmartOS specific output.Well thats a Bosh feature after all.. Cloud independant…

Patch the micro and cpi plugin

As of this writing, the latest bosh version is 1.2671.0. This version need to be changed in a few places to make bosh actually work with the external cpi. Discussion here

The diff is here

What i actually do, is clone the bosh repo, apply that diff, and:

$ export VER="1.2671.0"
$ export BOSH_REPO="/Users/killfill/Public/bosh-master"
$ cd ~/.rvm/gems/ruby-1.9.3-p484/gems
$ LIST=bosh_cli_plugin_micro bosh_cpi
$ for i in $LIST; do mv $i-$VER $i-$VER.orig; ln -s $BOSH_REPO/$i $i-$VER; done

If anyone is actually trying these changes, successfully or not, please let me know!

Conclusion

  1. Its possible to integrate Bosh with any Cloud infrastructure that manages Linux VMs and provides an API using external CPI like was described here.
  2. Using an external CPI on Bosh require some modifications, it is not fully supported yet, but hopefully will be in the near future.
  3. Cloud Foundry runs on FiFo / SmartOS just fine
  4. Bosh should be capable of handling deployments of different Operating Systems.
  5. It may be not very exiting to deploy Linux VMs for the SmartOS fan, but with some modifications it should handle SmartOS zones just fine.

⋅ DevOps ⋅ 2 comments