FAQ
Hi folks,

We released a lightweight (or restricted) BOSH job/package installer.

* https://github.com/nttlabs/nise_bosh

You can install jobs/packages included in cf-release repository with this
installer on your existing servers.
Since it does not depend on any CPI and it just runs the packaging script
for the given job on the server, you do not need any IaaS layer support to
use well-configured jobs/packages in cf-release.
We are using this installer to test components which are not supported by
vcap_dev_setup script, such as DEA/NG and CC/NG.

The scripts included in the installer do not completely emulate the
original BOSH's behavior, so some jobs/packages could not work properly.
Any feed backs and bug reports are welcome.

Search Discussions

  • Dr Nic Williams at Mar 11, 2013 at 10:10 am
    This very cool seeing more bosh/cf-release by-products being created as people figure out new ways to use the core cf-release asset!
    --
    Dr Nic Williams
    Stark & Wayne LLC - the consultancy for Cloud Foundry
    http://starkandwayne.com
    +1 415 860 2185
    twitter: drnic
    On Mon, Mar 11, 2013 at 4:51 AM, Iwasaki Yudai wrote:

    Hi folks,
    We released a lightweight (or restricted) BOSH job/package installer.
    * https://github.com/nttlabs/nise_bosh
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Since it does not depend on any CPI and it just runs the packaging script
    for the given job on the server, you do not need any IaaS layer support to
    use well-configured jobs/packages in cf-release.
    We are using this installer to test components which are not supported by
    vcap_dev_setup script, such as DEA/NG and CC/NG.
    The scripts included in the installer do not completely emulate the
    original BOSH's behavior, so some jobs/packages could not work properly.
    Any feed backs and bug reports are welcome.
  • Iwasaki Yudai at May 10, 2013 at 8:47 am
    I've released the Nise BOSH version 2.

    * https://github.com/nttlabs/nise_bosh

    It was rewritten to introduce BOSH's codebase and supports job collocation
    and 'properties' definitions in job template spec files.

    You can create a devbox (like MCF) with the new Nise BOSH with the manifest
    below (base on micro/deploy/manifest-ng.yml):

    https://gist.github.com/yudai/5553213

    Run `nise-bosh` command twice to install:
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh ~/cf-release ~/micro.yml
    micro
    sudo env PATH=$PATH bundle exec --keep-monit-files ./bin/nise-bosh
    ~/cf-release ~/micro.yml micro_ng

    See the README.md for the detail.
    On Monday, March 11, 2013 7:10:44 PM UTC+9, Dr Nic Williams wrote:

    This very cool seeing more bosh/cf-release by-products being created as
    people figure out new ways to use the core cf-release asset!
    --
    Dr Nic Williams
    Stark & Wayne LLC - the consultancy for Cloud Foundry
    http://starkandwayne.com
    +1 415 860 2185
    twitter: drnic


    On Mon, Mar 11, 2013 at 4:51 AM, Iwasaki Yudai <iwasak...@gmail.com<javascript:>
    wrote:
    Hi folks,

    We released a lightweight (or restricted) BOSH job/package installer.

    * https://github.com/nttlabs/nise_bosh

    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Since it does not depend on any CPI and it just runs the packaging script
    for the given job on the server, you do not need any IaaS layer support to
    use well-configured jobs/packages in cf-release.
    We are using this installer to test components which are not supported by
    vcap_dev_setup script, such as DEA/NG and CC/NG.

    The scripts included in the installer do not completely emulate the
    original BOSH's behavior, so some jobs/packages could not work properly.
    Any feed backs and bug reports are welcome.
  • Dr Nic Williams at May 10, 2013 at 5:51 pm
    nise!


    On Fri, May 10, 2013 at 1:47 AM, Iwasaki Yudai
    wrote:
    I've released the Nise BOSH version 2.

    * https://github.com/nttlabs/nise_bosh

    It was rewritten to introduce BOSH's codebase and supports job collocation
    and 'properties' definitions in job template spec files.

    You can create a devbox (like MCF) with the new Nise BOSH with the
    manifest below (base on micro/deploy/manifest-ng.yml):

    https://gist.github.com/yudai/5553213

    Run `nise-bosh` command twice to install:
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh ~/cf-release ~/micro.yml
    micro
    sudo env PATH=$PATH bundle exec --keep-monit-files ./bin/nise-bosh
    ~/cf-release ~/micro.yml micro_ng

    See the README.md for the detail.
    On Monday, March 11, 2013 7:10:44 PM UTC+9, Dr Nic Williams wrote:

    This very cool seeing more bosh/cf-release by-products being created as
    people figure out new ways to use the core cf-release asset!
    --
    Dr Nic Williams
    Stark & Wayne LLC - the consultancy for Cloud Foundry
    http://starkandwayne.com
    +1 415 860 2185
    twitter: drnic

    On Mon, Mar 11, 2013 at 4:51 AM, Iwasaki Yudai wrote:

    Hi folks,

    We released a lightweight (or restricted) BOSH job/package installer.

    * https://github.com/nttlabs/**nise_bosh<https://github.com/nttlabs/nise_bosh>

    You can install jobs/packages included in cf-release repository with
    this installer on your existing servers.
    Since it does not depend on any CPI and it just runs the packaging
    script for the given job on the server, you do not need any IaaS layer
    support to use well-configured jobs/packages in cf-release.
    We are using this installer to test components which are not supported
    by vcap_dev_setup script, such as DEA/NG and CC/NG.

    The scripts included in the installer do not completely emulate the
    original BOSH's behavior, so some jobs/packages could not work properly.
    Any feed backs and bug reports are welcome.

    --
    Dr Nic Williams
    Stark & Wayne LLC - consultancy for Cloud Foundry users
    http://drnicwilliams.com
    http://starkandwayne.com
    cell +1 (415) 860-2185
    twitter @drnic
  • Brian McClain at May 10, 2013 at 7:25 pm
    VERY cool! Saw the PR to add colocation, excited to see the feature's in
    place. Firing up a Vagrant machine as I type this :)

    - Brian
    @BrianMMcClain
    On Friday, May 10, 2013 4:47:53 AM UTC-4, Iwasaki Yudai wrote:

    I've released the Nise BOSH version 2.

    * https://github.com/nttlabs/nise_bosh

    It was rewritten to introduce BOSH's codebase and supports job collocation
    and 'properties' definitions in job template spec files.

    You can create a devbox (like MCF) with the new Nise BOSH with the
    manifest below (base on micro/deploy/manifest-ng.yml):

    https://gist.github.com/yudai/5553213

    Run `nise-bosh` command twice to install:
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh ~/cf-release ~/micro.yml
    micro
    sudo env PATH=$PATH bundle exec --keep-monit-files ./bin/nise-bosh
    ~/cf-release ~/micro.yml micro_ng

    See the README.md for the detail.
    On Monday, March 11, 2013 7:10:44 PM UTC+9, Dr Nic Williams wrote:

    This very cool seeing more bosh/cf-release by-products being created as
    people figure out new ways to use the core cf-release asset!
    --
    Dr Nic Williams
    Stark & Wayne LLC - the consultancy for Cloud Foundry
    http://starkandwayne.com
    +1 415 860 2185
    twitter: drnic

    On Mon, Mar 11, 2013 at 4:51 AM, Iwasaki Yudai wrote:

    Hi folks,

    We released a lightweight (or restricted) BOSH job/package installer.

    * https://github.com/nttlabs/nise_bosh

    You can install jobs/packages included in cf-release repository with
    this installer on your existing servers.
    Since it does not depend on any CPI and it just runs the packaging
    script for the given job on the server, you do not need any IaaS layer
    support to use well-configured jobs/packages in cf-release.
    We are using this installer to test components which are not supported
    by vcap_dev_setup script, such as DEA/NG and CC/NG.

    The scripts included in the installer do not completely emulate the
    original BOSH's behavior, so some jobs/packages could not work properly.
    Any feed backs and bug reports are welcome.
  • Gandalf Corvotempesta at May 10, 2013 at 9:08 pm

    2013/3/11 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Iwasaki Yudai at May 11, 2013 at 2:11 pm
    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a single
    server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Cbussler at May 13, 2013 at 4:59 pm
    Hi,

    just for clarification, in the case of a multi-server install, can this
    directly install on the server operating system ('bare metal') or does it
    need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Iwasaki Yudai at May 14, 2013 at 1:58 am
    Nise BOSH installs CF components on pre-installed Ubuntu servers. You don't
    need any IaaS layers to use Nise BOSH (if you have an IaaS infrastructure,
    I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to run
    Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can this
    directly install on the server operating system ('bare metal') or does it
    need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Iwasaki Yudai at May 14, 2013 at 6:44 am
    I created a install script for devbox. It works with Ubuntu 10.04 64bit.

    * https://gist.github.com/yudai/5553213

    You can login to your CF instance with micro@vcap.me / micr0@micr0 .
    (https://github.com/cloudfoundry/cf-release/commit/baaef065bab706e5deca37cf447ba3181731bec0
    made short passwords such as 'micro' prohibited. This change might make
    MCF-ng broken too?)

    Use admin-cf-plugin to set auth tokens and make services workable.

    * https://github.com/cloudfoundry/admin-cf-plugin



    On Tuesday, May 14, 2013 10:58:41 AM UTC+9, Iwasaki Yudai wrote:

    Nise BOSH installs CF components on pre-installed Ubuntu servers. You
    don't need any IaaS layers to use Nise BOSH (if you have an IaaS
    infrastructure, I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to
    run Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can this
    directly install on the server operating system ('bare metal') or does it
    need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Matt Reider at May 14, 2013 at 7:21 am
    Excited to try this tomorrow.

    Sent from my iPhone
    On May 13, 2013, at 11:44 PM, Iwasaki Yudai wrote:

    I created a install script for devbox. It works with Ubuntu 10.04 64bit.

    * https://gist.github.com/yudai/5553213

    You can login to your CF instance with micro@vcap.me / micr0@micr0 .
    (https://github.com/cloudfoundry/cf-release/commit/baaef065bab706e5deca37cf447ba3181731bec0 made short passwords such as 'micro' prohibited. This change might make MCF-ng broken too?)

    Use admin-cf-plugin to set auth tokens and make services workable.

    * https://github.com/cloudfoundry/admin-cf-plugin



    On Tuesday, May 14, 2013 10:58:41 AM UTC+9, Iwasaki Yudai wrote:

    Nise BOSH installs CF components on pre-installed Ubuntu servers. You don't need any IaaS layers to use Nise BOSH (if you have an IaaS infrastructure, I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to run Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can this directly install on the server operating system ('bare metal') or does it need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a single server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • David Dauer at May 15, 2013 at 7:05 am
    Great work! Could you also let us know how to proceed in case we want to
    update a devbox installed with this script?
    On Tuesday, May 14, 2013 8:44:14 AM UTC+2, Iwasaki Yudai wrote:

    I created a install script for devbox. It works with Ubuntu 10.04 64bit.

    * https://gist.github.com/yudai/5553213

    You can login to your CF instance with mi...@vcap.me <javascript:> /
    micr0@micr0 .
    (
    https://github.com/cloudfoundry/cf-release/commit/baaef065bab706e5deca37cf447ba3181731bec0made short passwords such as 'micro' prohibited. This change might make
    MCF-ng broken too?)

    Use admin-cf-plugin to set auth tokens and make services workable.

    * https://github.com/cloudfoundry/admin-cf-plugin



    On Tuesday, May 14, 2013 10:58:41 AM UTC+9, Iwasaki Yudai wrote:

    Nise BOSH installs CF components on pre-installed Ubuntu servers. You
    don't need any IaaS layers to use Nise BOSH (if you have an IaaS
    infrastructure, I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to
    run Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can this
    directly install on the server operating system ('bare metal') or does it
    need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.
    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Yudai Iwasaki at May 15, 2013 at 7:13 am
    You can update your instance by the following steps:

    1. recreate a new release

    cd ~/cf-release
    # do something to update your release
    bosh create release --force

    2. stop all job processes once

    sudo/var/vcap/bosh/bin/monit stop all

    3. modify your deployment manifest if needed

    vi ~/micro.yml


    4. rerun Nise BOSH

    cd ~/nise-bosh
    sudo env PATH=$PATH bundle install./bin/nise-bosh ~/cf-release ~/micro.yml
    micro
    sudo env PATH=$PATH bundle install./bin/nise-bosh --keep-monit-files ~/cf-release
    ~/micro.yml micro_ng

    5. launch processes

    sudo/var/vcap/bosh/bin/monit start all

    Have fun!
    On Wednesday, May 15, 2013 4:04:59 PM UTC+9, David Dauer wrote:

    Great work! Could you also let us know how to proceed in case we want to
    update a devbox installed with this script?
    On Tuesday, May 14, 2013 8:44:14 AM UTC+2, Iwasaki Yudai wrote:

    I created a install script for devbox. It works with Ubuntu 10.04 64bit.

    * https://gist.github.com/yudai/5553213

    You can login to your CF instance with mi...@vcap.me / micr0@micr0 .
    (
    https://github.com/cloudfoundry/cf-release/commit/baaef065bab706e5deca37cf447ba3181731bec0made short passwords such as 'micro' prohibited. This change might make
    MCF-ng broken too?)

    Use admin-cf-plugin to set auth tokens and make services workable.

    * https://github.com/cloudfoundry/admin-cf-plugin



    On Tuesday, May 14, 2013 10:58:41 AM UTC+9, Iwasaki Yudai wrote:

    Nise BOSH installs CF components on pre-installed Ubuntu servers. You
    don't need any IaaS layers to use Nise BOSH (if you have an IaaS
    infrastructure, I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to
    run Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can this
    directly install on the server operating system ('bare metal') or does it
    need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.

    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta
    wrote:
    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Yudai Iwasaki at May 15, 2013 at 7:16 am
    Oops, in the step 4, replace two lines please;

    s/bundle install./bundle exec ./g
    On Wednesday, May 15, 2013 4:13:33 PM UTC+9, Yudai Iwasaki wrote:

    You can update your instance by the following steps:

    1. recreate a new release

    cd ~/cf-release
    # do something to update your release
    bosh create release --force

    2. stop all job processes once

    sudo/var/vcap/bosh/bin/monit stop all

    3. modify your deployment manifest if needed

    vi ~/micro.yml


    4. rerun Nise BOSH

    cd ~/nise-bosh
    sudo env PATH=$PATH bundle install./bin/nise-bosh ~/cf-release ~/micro.yml
    micro
    sudo env PATH=$PATH bundle install./bin/nise-bosh --keep-monit-files ~/cf-release
    ~/micro.yml micro_ng

    5. launch processes

    sudo/var/vcap/bosh/bin/monit start all

    Have fun!
    On Wednesday, May 15, 2013 4:04:59 PM UTC+9, David Dauer wrote:

    Great work! Could you also let us know how to proceed in case we want to
    update a devbox installed with this script?
    On Tuesday, May 14, 2013 8:44:14 AM UTC+2, Iwasaki Yudai wrote:

    I created a install script for devbox. It works with Ubuntu 10.04 64bit.

    * https://gist.github.com/yudai/5553213

    You can login to your CF instance with mi...@vcap.me / micr0@micr0 .
    (
    https://github.com/cloudfoundry/cf-release/commit/baaef065bab706e5deca37cf447ba3181731bec0made short passwords such as 'micro' prohibited. This change might make
    MCF-ng broken too?)

    Use admin-cf-plugin to set auth tokens and make services workable.

    * https://github.com/cloudfoundry/admin-cf-plugin



    On Tuesday, May 14, 2013 10:58:41 AM UTC+9, Iwasaki Yudai wrote:

    Nise BOSH installs CF components on pre-installed Ubuntu servers. You
    don't need any IaaS layers to use Nise BOSH (if you have an IaaS
    infrastructure, I would recommend to use BOSH).
    When you want to install CF components on multiple servers, you need to
    run Nise BOSH on each of the servers. You can use Capistrano to do so.
    On Tuesday, May 14, 2013 1:59:20 AM UTC+9, cbus...@voxeolabs.com wrote:

    Hi,

    just for clarification, in the case of a multi-server install, can
    this directly install on the server operating system ('bare metal') or does
    it need a virtual machine infrastructure?

    Thanks,

    Christoph
    On Saturday, May 11, 2013 7:11:14 AM UTC-7, Iwasaki Yudai wrote:

    You can install everything both on a single node and multiple servers.

    I'll write some chef specs or scripts to install all components on a
    single server with a single command.

    On Saturday, May 11, 2013 6:07:57 AM UTC+9, Gandalf Corvotempesta
    wrote:
    2013/3/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install jobs/packages included in cf-release repository with this
    installer on your existing servers.
    Is this multiserver or it will install everything on a single node?
  • Gandalf Corvotempesta at May 15, 2013 at 7:26 am

    2013/5/11 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    You can install everything both on a single node and multiple servers.
    But doing this require a manual install over each server, right?
    Actually is not possible to use Nise Bosh like the real bosh starting
    from a workstation and then updating the whole infrastructure across
    multiple servers?

    Just to better understand.
  • Yudai Iwasaki at May 15, 2013 at 7:57 am
    Yes, your understanding is exactly correct. I'm sorry my wording was
    ambiguous.
    Nise BOSH itself does not has functionality to configurate multiple servers.
    You need to write some Capistrano tasks to do so.
    On Wednesday, May 15, 2013 4:26:27 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/11 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    You can install everything both on a single node and multiple servers.
    But doing this require a manual install over each server, right?
    Actually is not possible to use Nise Bosh like the real bosh starting
    from a workstation and then updating the whole infrastructure across
    multiple servers?

    Just to better understand.
  • Brian McClain at May 16, 2013 at 1:13 am
    By the way, I spun up a machine with this and it worked like a charm.
    Amazing work on this! First time I've had a working CF v2 system up and
    running, and I've been giggling like a child all night

    FWIW, I might have have overlooked something, or might just not know how
    the bootstrap admin account works, but if it helps anyone else, the login
    after CF was up and running was the following:

    Email: micro@vcap.me
    Password: micr0@micr0

    - Brian
    On Wednesday, May 15, 2013 3:57:46 AM UTC-4, Iwasaki Yudai wrote:

    Yes, your understanding is exactly correct. I'm sorry my wording was
    ambiguous.
    Nise BOSH itself does not has functionality to configurate multiple
    servers.
    You need to write some Capistrano tasks to do so.
    On Wednesday, May 15, 2013 4:26:27 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install everything both on a single node and multiple servers.
    But doing this require a manual install over each server, right?
    Actually is not possible to use Nise Bosh like the real bosh starting
    from a workstation and then updating the whole infrastructure across
    multiple servers?

    Just to better understand.
  • Matt Reider at May 16, 2013 at 5:43 am
    Yep. Huge kudos. This is a real milestone.

    Sent from my iPhone
    On May 15, 2013, at 6:13 PM, Brian McClain wrote:

    By the way, I spun up a machine with this and it worked like a charm. Amazing work on this! First time I've had a working CF v2 system up and running, and I've been giggling like a child all night

    FWIW, I might have have overlooked something, or might just not know how the bootstrap admin account works, but if it helps anyone else, the login after CF was up and running was the following:

    Email: micro@vcap.me
    Password: micr0@micr0

    - Brian
    On Wednesday, May 15, 2013 3:57:46 AM UTC-4, Iwasaki Yudai wrote:

    Yes, your understanding is exactly correct. I'm sorry my wording was ambiguous.
    Nise BOSH itself does not has functionality to configurate multiple servers.
    You need to write some Capistrano tasks to do so.
    On Wednesday, May 15, 2013 4:26:27 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/11 Iwasaki Yudai <iwasak...@gmail.com>:
    You can install everything both on a single node and multiple servers.
    But doing this require a manual install over each server, right?
    Actually is not possible to use Nise Bosh like the real bosh starting
    from a workstation and then updating the whole infrastructure across
    multiple servers?

    Just to better understand.
  • Gandalf Corvotempesta at May 16, 2013 at 7:35 am

    2013/5/16 Brian McClain <brianmmcclain@gmail.com>:
    By the way, I spun up a machine with this and it worked like a charm.
    Amazing work on this! First time I've had a working CF v2 system up and
    running, and I've been giggling like a child all night
    I would like to test this system.
    The only thing that I have to run is this:
    https://gist.github.com/yudai/5553213
    right?

    And if I would like to try with multiple server, should I run the same
    script buy changing micro.yml according to my configuration?

    If this is good, this will be the better system to use to install CF,
    even in production.
  • Iwasaki Yudai at May 16, 2013 at 7:49 am
    Basically yes.

    Modify the manifest file as you like and run Nise BOSH on each server.
    You can separate 'micro' and 'micro_ng' job into multiple jobs if needed,
    like:

    jobs:
       - name: for_server1
         template:
            - nats
       - name: for_server2
         template:
            - dea_next
    ...

    You can also use the "-i" option to set index property for each server.
    This is useful when use install the same job template into multiple servers.

    On Thursday, May 16, 2013 4:35:17 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/16 Brian McClain <brianm...@gmail.com <javascript:>>:
    By the way, I spun up a machine with this and it worked like a charm.
    Amazing work on this! First time I've had a working CF v2 system up and
    running, and I've been giggling like a child all night
    I would like to test this system.
    The only thing that I have to run is this:
    https://gist.github.com/yudai/5553213
    right?

    And if I would like to try with multiple server, should I run the same
    script buy changing micro.yml according to my configuration?

    If this is good, this will be the better system to use to install CF,
    even in production.
  • Gandalf Corvotempesta at May 16, 2013 at 8:14 am

    2013/5/16 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Modify the manifest file as you like and run Nise BOSH on each server.
    You can separate 'micro' and 'micro_ng' job into multiple jobs if needed,
    like:

    jobs:
    - name: for_server1
    template:
    - nats
    - name: for_server2
    template:
    - dea_next
    ...

    You can also use the "-i" option to set index property for each server. This
    is useful when use install the same job template into multiple servers.
    Great, in this way i'll be able to use a single file everywhere.
    Putting this under a local git will be easy to script a whole
    infrastructure installation and sync everything.

    I'll try immediatly.
  • Gandalf Corvotempesta at May 16, 2013 at 3:25 pm

    2013/5/16 Gandalf Corvotempesta <gandalf.corvotempesta@gmail.com>:
    I would like to test this system.
    The only thing that I have to run is this:
    https://gist.github.com/yudai/5553213
    right?
    After a succesfultl installation i'm unable to get mysql_worker,
    postgresql_worker and gorouter up and running:

    # sudo /var/vcap/bosh/bin/monit start all
    'postgres' start: /var/vcap/jobs/postgres/bin/postgres_ctl
    'nats' start: /var/vcap/jobs/nats/bin/nats_ctl
    'warden' start: /var/vcap/jobs/dea_next/bin/warden_ctl
    'dea_next' start: /var/vcap/jobs/dea_next/bin/dea_ctl
    'dir_server' start: /var/vcap/jobs/dea_next/bin/dir_server_ctl
    'health_manager_next' start:
    /var/vcap/jobs/health_manager_next/bin/health_manager_next_ctl
    'serialization_data_server' start:
    /var/vcap/jobs/serialization_data_server/bin/serialization_data_server_ctl
    'uaa' start: /var/vcap/jobs/uaa/bin/uaa_ctl
    'vcap_registrar' start: /var/vcap/jobs/uaa/bin/vcap_registrar_ctl
    'vcap_redis' start: /var/vcap/jobs/vcap_redis/bin/vcap_redis_ctl
    'mysql55' start: /var/vcap/jobs/mysql_node_ng/bin/mysql55_ctl
    'mysql_node' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_node_ctl
    'mysql_worker_0' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_0' failed to start
    'mysql_worker_1' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_1' failed to start
    'mysql_worker_2' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_2' failed to start
    'mysql_worker_3' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_3' failed to start
    'mysql_worker_4' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_4' failed to start
    'mysql_gateway' start: /var/vcap/jobs/mysql_gateway/bin/mysql_gateway_ctl
    'cloud_controller_ng' start:
    /var/vcap/jobs/cloud_controller_ng/bin/cloud_controller_ng_ctl
    'postgresql91' start: /var/vcap/jobs/postgresql_node_ng/bin/postgresql_ctl
    'postgresql_node' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_node_ctl
    'postgresql_worker_0' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_0' failed to start
    'postgresql_worker_1' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_1' failed to start
    'postgresql_worker_2' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_2' failed to start
    'postgresql_worker_3' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_3' failed to start
    'postgresql_worker_4' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_4' failed to start
    'gorouter' start: /var/vcap/jobs/gorouter/bin/gorouter_ctl
    'gorouter' failed to start
    'postgresql_gateway' start:
    /var/vcap/jobs/postgresql_gateway/bin/postgresql_gateway_ctl
  • Iwasaki Yudai at May 17, 2013 at 2:28 am
    Thanks a lot for your feedback!
    I've updated install.sh and introduced your suggestions.
    However, I could not figure out which command makes the error 'typeset: not
    found'. Could you show me the line number or command? Since install.sh has
    #!/bin/bash at the beginning, it might be raised in a child process.

    I confirmed sometimes some processes fail to start at the first run. I'm
    not sure why this happens, but it may be because of the order of starting
    up the jobs because some jobs takes longer time to initialize their
    environments at the first run.
    I recommend you to use a Monit daemon to retry launching jobs. You can
    start a monit daemon by executing the 'sudo /var/vcap/bosh/bin/monit'
    command with no options. Or, you can manually retry starting up by
    repeating the 'start up' command.

    If the problem continues, you can examine the cause by looking into logs in
    /var/vcap/sys/log (root privilege required). Logs in
    /var/vcap/sys/log/monit is especially helpful for trouble shooting.
    On Friday, May 17, 2013 12:25:40 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/16 Gandalf Corvotempesta <gandalf.co...@gmail.com <javascript:>>:
    I would like to test this system.
    The only thing that I have to run is this:
    https://gist.github.com/yudai/5553213
    right?
    After a succesfultl installation i'm unable to get mysql_worker,
    postgresql_worker and gorouter up and running:

    # sudo /var/vcap/bosh/bin/monit start all
    'postgres' start: /var/vcap/jobs/postgres/bin/postgres_ctl
    'nats' start: /var/vcap/jobs/nats/bin/nats_ctl
    'warden' start: /var/vcap/jobs/dea_next/bin/warden_ctl
    'dea_next' start: /var/vcap/jobs/dea_next/bin/dea_ctl
    'dir_server' start: /var/vcap/jobs/dea_next/bin/dir_server_ctl
    'health_manager_next' start:
    /var/vcap/jobs/health_manager_next/bin/health_manager_next_ctl
    'serialization_data_server' start:
    /var/vcap/jobs/serialization_data_server/bin/serialization_data_server_ctl
    'uaa' start: /var/vcap/jobs/uaa/bin/uaa_ctl
    'vcap_registrar' start: /var/vcap/jobs/uaa/bin/vcap_registrar_ctl
    'vcap_redis' start: /var/vcap/jobs/vcap_redis/bin/vcap_redis_ctl
    'mysql55' start: /var/vcap/jobs/mysql_node_ng/bin/mysql55_ctl
    'mysql_node' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_node_ctl
    'mysql_worker_0' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_0' failed to start
    'mysql_worker_1' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_1' failed to start
    'mysql_worker_2' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_2' failed to start
    'mysql_worker_3' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_3' failed to start
    'mysql_worker_4' start: /var/vcap/jobs/mysql_node_ng/bin/mysql_worker_ctl
    'mysql_worker_4' failed to start
    'mysql_gateway' start: /var/vcap/jobs/mysql_gateway/bin/mysql_gateway_ctl
    'cloud_controller_ng' start:
    /var/vcap/jobs/cloud_controller_ng/bin/cloud_controller_ng_ctl
    'postgresql91' start: /var/vcap/jobs/postgresql_node_ng/bin/postgresql_ctl
    'postgresql_node' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_node_ctl
    'postgresql_worker_0' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_0' failed to start
    'postgresql_worker_1' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_1' failed to start
    'postgresql_worker_2' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_2' failed to start
    'postgresql_worker_3' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_3' failed to start
    'postgresql_worker_4' start:
    /var/vcap/jobs/postgresql_node_ng/bin/postgresql_worker_ctl
    'postgresql_worker_4' failed to start
    'gorouter' start: /var/vcap/jobs/gorouter/bin/gorouter_ctl
    'gorouter' failed to start
    'postgresql_gateway' start:
    /var/vcap/jobs/postgresql_gateway/bin/postgresql_gateway_ctl
  • Gandalf Corvotempesta at May 17, 2013 at 7:06 am
    In Ubuntu default shell is dash and bash is usually linked to dash.
    Changing dash to dash will solve the issue.

    I've tried many times to start all processes with no success. My SQL
    doesn't start. It seems an issue related to some deprecated function in
    Ruby and it also trying to connect to redis service.

    I'll post some log in next hour
    Il giorno 17/mag/2013 04:28, "Iwasaki Yudai" <iwasaki.yudai@gmail.com> ha
    scritto:
    Thanks a lot for your feedback!
    I've updated install.sh and introduced your suggestions.
    However, I could not figure out which command makes the error 'typeset:
    not found'. Could you show me the line number or command? Since install.sh
    has #!/bin/bash at the beginning, it might be raised in a child process.

    I confirmed sometimes some processes fail to start at the first run. I'm
    not sure why this happens, but it may be because of the order of starting
    up the jobs because some jobs takes longer time to initialize their
    environments at the first run.
    I recommend you to use a Monit daemon to retry launching jobs. You can
    start a monit daemon by executing the 'sudo /var/vcap/bosh/bin/monit'
    command with no options. Or, you can manually retry starting up by
    repeating the 'start up' command.

    If the problem continues, you can examine the cause by looking into logs
    in /var/vcap/sys/log (root privilege required). Logs in
    /var/vcap/sys/log/monit is especially helpful for trouble shooting.
    On Friday, May 17, 2013 12:25:40 AM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/16 Gandalf Corvotempesta <gandalf.co...@gmail.**com>:
    I would like to test this system.
    The only thing that I have to run is this:
    https://gist.github.com/yudai/**5553213<https://gist.github.com/yudai/5553213>
    right?
    After a succesfultl installation i'm unable to get mysql_worker,
    postgresql_worker and gorouter up and running:

    # sudo /var/vcap/bosh/bin/monit start all
    'postgres' start: /var/vcap/jobs/postgres/bin/**postgres_ctl
    'nats' start: /var/vcap/jobs/nats/bin/nats_**ctl
    'warden' start: /var/vcap/jobs/dea_next/bin/**warden_ctl
    'dea_next' start: /var/vcap/jobs/dea_next/bin/**dea_ctl
    'dir_server' start: /var/vcap/jobs/dea_next/bin/**dir_server_ctl
    'health_manager_next' start:
    /var/vcap/jobs/health_manager_**next/bin/health_manager_next_**ctl
    'serialization_data_server' start:
    /var/vcap/jobs/serialization_**data_server/bin/serialization_**data_server_ctl

    'uaa' start: /var/vcap/jobs/uaa/bin/uaa_ctl
    'vcap_registrar' start: /var/vcap/jobs/uaa/bin/vcap_**registrar_ctl
    'vcap_redis' start: /var/vcap/jobs/vcap_redis/bin/**vcap_redis_ctl
    'mysql55' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql55_ctl
    'mysql_node' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_node_ctl
    'mysql_worker_0' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_worker_ctl

    'mysql_worker_0' failed to start
    'mysql_worker_1' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_worker_ctl

    'mysql_worker_1' failed to start
    'mysql_worker_2' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_worker_ctl

    'mysql_worker_2' failed to start
    'mysql_worker_3' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_worker_ctl

    'mysql_worker_3' failed to start
    'mysql_worker_4' start: /var/vcap/jobs/mysql_node_ng/**bin/mysql_worker_ctl

    'mysql_worker_4' failed to start
    'mysql_gateway' start: /var/vcap/jobs/mysql_gateway/**bin/mysql_gateway_ctl

    'cloud_controller_ng' start:
    /var/vcap/jobs/cloud_**controller_ng/bin/cloud_**controller_ng_ctl
    'postgresql91' start: /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_ctl

    'postgresql_node' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_node_**ctl
    'postgresql_worker_0' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_worker_**ctl
    'postgresql_worker_0' failed to start
    'postgresql_worker_1' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_worker_**ctl
    'postgresql_worker_1' failed to start
    'postgresql_worker_2' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_worker_**ctl
    'postgresql_worker_2' failed to start
    'postgresql_worker_3' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_worker_**ctl
    'postgresql_worker_3' failed to start
    'postgresql_worker_4' start:
    /var/vcap/jobs/postgresql_**node_ng/bin/postgresql_worker_**ctl
    'postgresql_worker_4' failed to start
    'gorouter' start: /var/vcap/jobs/gorouter/bin/**gorouter_ctl
    'gorouter' failed to start
    'postgresql_gateway' start:
    /var/vcap/jobs/postgresql_**gateway/bin/postgresql_**gateway_ctl
  • Gandalf Corvotempesta at May 17, 2013 at 8:59 am

    2013/5/17 Gandalf Corvotempesta <gandalf.corvotempesta@gmail.com>:
    I'll post some log in next hour
    this is mysql_node_worker log:

    http://pastebin.com/raw.php?i=vmfP38PX
  • Gandalf Corvotempesta at May 17, 2013 at 9:00 am

    2013/5/17 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    I confirmed sometimes some processes fail to start at the first run. I'm not
    sure why this happens, but it may be because of the order of starting up the
    jobs because some jobs takes longer time to initialize their environments at
    the first run.
    I recommend you to use a Monit daemon to retry launching jobs. You can start
    a monit daemon by executing the 'sudo /var/vcap/bosh/bin/monit' command with
    no options. Or, you can manually retry starting up by repeating the 'start
    up' command.
    Monit is stuck here:
    # sudo /var/vcap/bosh/bin/monit status
    The Monit daemon 5.2.4 uptime: 3m

    Process 'postgres'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'nats'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'warden'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'dea_next'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'dir_server'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'health_manager_next'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'serialization_data_server'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'uaa'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'vcap_registrar'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'vcap_redis'
       status initializing
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql55'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_node'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_0'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_1'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_2'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_3'
       status initializing - start pending
       monitoring status initializing
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_4'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'mysql_gateway'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'cloud_controller_ng'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'nginx_ccng'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql91'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_node'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_0'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_1'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_2'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_3'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_4'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'gorouter'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    Process 'postgresql_gateway'
       status not monitored - start pending
       monitoring status not monitored
       data collected Fri May 17 10:56:31 2013

    System 'system_microcloudfoundry'
       status running - start pending
       monitoring status monitored
       load average [1.35] [0.59] [0.22]
       cpu 0.7%us 0.6%sy 0.3%wa
       memory usage 47348 kB [4.6%]
       swap usage 0 kB [0.0%]
       data collected Fri May 17 10:56:31 2013
  • Iwasaki Yudai at May 17, 2013 at 9:09 am
    It seems all the processes are still initializing. On my environment, it
    takes about 6 minutes to complete startup.
    Could you show me the monit status after waiting for about 10 minutes?
    On Friday, May 17, 2013 6:00:18 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/17 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    I confirmed sometimes some processes fail to start at the first run. I'm not
    sure why this happens, but it may be because of the order of starting up the
    jobs because some jobs takes longer time to initialize their
    environments at
    the first run.
    I recommend you to use a Monit daemon to retry launching jobs. You can start
    a monit daemon by executing the 'sudo /var/vcap/bosh/bin/monit' command with
    no options. Or, you can manually retry starting up by repeating the 'start
    up' command.
    Monit is stuck here:
    # sudo /var/vcap/bosh/bin/monit status
    The Monit daemon 5.2.4 uptime: 3m

    Process 'postgres'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'nats'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'warden'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'dea_next'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'dir_server'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'health_manager_next'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'serialization_data_server'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'uaa'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'vcap_registrar'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'vcap_redis'
    status initializing
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql55'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_node'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_0'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_1'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_2'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_3'
    status initializing - start pending
    monitoring status initializing
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_worker_4'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'mysql_gateway'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'cloud_controller_ng'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'nginx_ccng'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql91'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_node'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_0'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_1'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_2'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_3'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_worker_4'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'gorouter'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    Process 'postgresql_gateway'
    status not monitored - start pending
    monitoring status not monitored
    data collected Fri May 17 10:56:31 2013

    System 'system_microcloudfoundry'
    status running - start pending
    monitoring status monitored
    load average [1.35] [0.59] [0.22]
    cpu 0.7%us 0.6%sy 0.3%wa
    memory usage 47348 kB [4.6%]
    swap usage 0 kB [0.0%]
    data collected Fri May 17 10:56:31 2013
  • Gandalf Corvotempesta at May 17, 2013 at 9:36 am

    2013/5/17 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    It seems all the processes are still initializing. On my environment, it
    takes about 6 minutes to complete startup.
    Could you show me the monit status after waiting for about 10 minutes?
    I'm using a VPS and it's very very slow. I'll try on a physical server.

    MySQL Workers should be real mysql server used by customers and placed
    multiple times on Micro CloudFoudry to simulate resource distribution,
    right? How can I decrease their number to 2? Having 5 MySQL and 5
    Postres on the same server is very heavy.
  • Gandalf Corvotempesta at May 17, 2013 at 1:07 pm

    2013/5/17 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Thanks a lot for your feedback!
    I've updated install.sh and introduced your suggestions.
    You have updated install.sh without proper quotes:

    current_ip=`ip addr | grep 'inet .*global' | cut -f 6 -d ' ' | cut -f1 -d '/'`

    inet .*global

    must be quoted, due to the space.
  • Iwasaki Yudai at May 17, 2013 at 1:33 pm
    Thanks! I fixed it.

    Workers are used to manage lifecycles. To my best knowledge, they are not
    so heavy themselves.
    Anyway, you can change the number of workers by editing 'worker_count' in
    each 'plans' property.
    (See
    also: https://github.com/cloudfoundry/cf-release/blob/master/jobs/postgresql_node_ng/monit#L66)
    On Friday, May 17, 2013 10:07:25 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/17 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    Thanks a lot for your feedback!
    I've updated install.sh and introduced your suggestions.
    You have updated install.sh without proper quotes:

    current_ip=`ip addr | grep 'inet .*global' | cut -f 6 -d ' ' | cut -f1 -d
    '/'`

    inet .*global

    must be quoted, due to the space.
  • Gandalf Corvotempesta at May 21, 2013 at 7:22 am

    2013/5/17 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Thanks! I fixed it.

    Workers are used to manage lifecycles. To my best knowledge, they are not so
    heavy themselves.
    Anyway, you can change the number of workers by editing 'worker_count' in
    each 'plans' property.
    (See also:
    https://github.com/cloudfoundry/cf-release/blob/master/jobs/postgresql_node_ng/monit#L66)

    Now i'm able to use this new installation as target for "cf" but there
    isn't any service provided.
    Is this normal? It's a blank installation?

    cf services

    doesn't show any available service.
  • Iwasaki Yudai at May 21, 2013 at 7:30 am
    Try 'cf info --services'. You need to create a organization (create-org)
    and a space (create-space, switch-space) before doing that.

    I did the same mistake as you a few days ago. It's a trap indeed.
    On Tuesday, May 21, 2013 4:16:06 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/17 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    Thanks! I fixed it.

    Workers are used to manage lifecycles. To my best knowledge, they are not so
    heavy themselves.
    Anyway, you can change the number of workers by editing 'worker_count' in
    each 'plans' property.
    (See also:
    https://github.com/cloudfoundry/cf-release/blob/master/jobs/postgresql_node_ng/monit#L66)



    Now i'm able to use this new installation as target for "cf" but there
    isn't any service provided.
    Is this normal? It's a blank installation?

    cf services

    doesn't show any available service.
  • Gandalf Corvotempesta at May 23, 2013 at 8:04 pm
    Only databases?

    $ cf info --services
    Getting services... OK

    service version provider plans description
    mysql 5.5 core 100 MySQL database
    postgresql 9.1 core 100 PostgreSQL database (vFabric)

    2013/5/21 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Try 'cf info --services'. You need to create a organization (create-org) and
    a space (create-space, switch-space) before doing that.
  • Gandalf Corvotempesta at May 23, 2013 at 9:08 pm

    2013/5/23 Gandalf Corvotempesta <gandalf.corvotempesta@gmail.com>:
    Only databases?

    $ cf info --services
    Getting services... OK

    service version provider plans description
    mysql 5.5 core 100 MySQL database
    postgresql 9.1 core 100 PostgreSQL database (vFabric)
    This is what I see in dea_next log:

    {"timestamp":1369339989.8300583,"message":"Connecting to NATS on
    nats://nats:nats@x:4222","log_level":"info","source":"Dea::Nats","data":{},"thread_id":18999420,"fiber_id":27446020,"process_id":21803,"file":"/var/vcap/packages/dea_next/lib/dea/nats.rb","lineno":95,"method":"create_nats_client"}
    {"timestamp":1369339992.782844,"message":"Received on
    \"router.start\": {\"id\"=>\"7bfa9dab89b3d833cbb4f4cb5c69c630\",
    \"hosts\"=>[\"x\"]}","log_level":"debug","source":"Dea::Nats","data":{},"thread_id":18999420,"fiber_id":27446020,"process_id":21803,"file":"/var/vcap/packages/dea_next/lib/dea/nats.rb","lineno":75,"method":"block
    in subscribe"}
    {"timestamp":1369340022.7827835,"message":"Received on
    \"router.start\": {\"id\"=>\"7bfa9dab89b3d833cbb4f4cb5c69c630\",
    \"hosts\"=>[\"x\"]}","log_level":"debug","source":"Dea::Nats","data":{},"thread_id":18999420,"fiber_id":27446020,"process_id":21803,"file":"/var/vcap/packages/dea_next/lib/dea/nats.rb","lineno":75,"method":"block
    in subscribe"}
  • Iwasaki Yudai at May 24, 2013 at 2:23 am
    Yes. My script currently installs only postgres and mysql.

    If you need any other services, you can add them by modifying your manifest.
    However, as far as I know, mongodb conflicted with something other job
    templates installed and did not work. It probably works if it is installed
    in an isolated VM.

    Which service especially do you need?
    On Friday, May 24, 2013 5:04:53 AM UTC+9, Gandalf Corvotempesta wrote:

    Only databases?

    $ cf info --services
    Getting services... OK

    service version provider plans description
    mysql 5.5 core 100 MySQL database
    postgresql 9.1 core 100 PostgreSQL database (vFabric)

    2013/5/21 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    Try 'cf info --services'. You need to create a organization (create-org) and
    a space (create-space, switch-space) before doing that.
  • Gandalf Corvotempesta at May 24, 2013 at 7:34 am

    2013/5/24 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Yes. My script currently installs only postgres and mysql.

    If you need any other services, you can add them by modifying your manifest.
    However, as far as I know, mongodb conflicted with something other job
    templates installed and did not work. It probably works if it is installed
    in an isolated VM.

    Which service especially do you need?
    I need a webserver. I've seen that dea_next and warden are both
    installed, up and running, but not available. MySQL and PGSQL are
    enough, but I also need a web.
  • Iwasaki Yudai at May 24, 2013 at 7:42 am
    Try 'cf info --all'. It shows other information on your CF instance.
    If you cannot push any applications, let me know.
    On Friday, May 24, 2013 4:34:32 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/24 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    Yes. My script currently installs only postgres and mysql.

    If you need any other services, you can add them by modifying your manifest.
    However, as far as I know, mongodb conflicted with something other job
    templates installed and did not work. It probably works if it is installed
    in an isolated VM.

    Which service especially do you need?
    I need a webserver. I've seen that dea_next and warden are both
    installed, up and running, but not available. MySQL and PGSQL are
    enough, but I also need a web.
  • Gandalf Corvotempesta at May 24, 2013 at 8:33 am

    2013/5/24 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Try 'cf info --all'. It shows other information on your CF instance.
    If you cannot push any applications, let me know.
    I'm reinstalling the whole system right now. It seems pretty unstable,
    sometimes services die and monit is unable to start them again.
    I have services in "initializing" state from yersteday.
  • Gandalf Corvotempesta at May 24, 2013 at 10:01 am

    2013/5/24 Gandalf Corvotempesta <gandalf.corvotempesta@gmail.com>:
    I'm reinstalling the whole system right now. It seems pretty unstable,
    sometimes services die and monit is unable to start them again.
    I have services in "initializing" state from yersteday.
    After a full reinstall, i'm unable to taget the new installation.
    It seems that no service is listening on port 80:

    # netstat -na4 | grep ":80 "
    #

    The Monit daemon 5.2.4 uptime: 10m

    Process 'postgres' Does not exist
    Process 'nats' running
    Process 'warden' running
    Process 'dea_next' running
    Process 'dir_server' running
    Process 'health_manager_next' running
    Process 'serialization_data_server' running
    Process 'uaa' running
    Process 'vcap_registrar' running
    Process 'vcap_redis' running
    Process 'mysql55' running
    Process 'mysql_node' running
    Process 'mysql_worker_0' running
    Process 'mysql_worker_1' running
    Process 'mysql_worker_2' running
    Process 'mysql_worker_3' running
    Process 'mysql_worker_4' running
    Process 'mysql_gateway' running
    Process 'cloud_controller_ng' running
    Process 'postgresql91' Does not exist
    Process 'postgresql_node' initializing
    Process 'postgresql_worker_0' not monitored
    Process 'postgresql_worker_1' not monitored
    Process 'postgresql_worker_2' not monitored
    Process 'postgresql_worker_3' not monitored
    Process 'postgresql_worker_4' not monitored
    Process 'gorouter' running
    Process 'postgresql_gateway' running
    System 'system_microcloudfoundry' running
  • Gandalf Corvotempesta at May 24, 2013 at 10:24 am

    2013/5/24 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    Try 'cf info --all'. It shows other information on your CF instance.
    If you cannot push any applications, let me know.
    $ cf info --all
    Getting services... OK

    Cloud Foundry sponsored by Pivotal

    target: http://api.vcap.me
       version: 2
       support: http://support.cloudfoundry.com

    user: micro@vcap.me

    service version provider plans description
    mysql 5.5 core 100 MySQL database
    postgresql 9.1 core 100 PostgreSQL database (vFabric)


    $ cf restart prova
    Using manifest file manifest.yml

    Stopping prova... OK

    Starting prova... OK
    -----> Downloaded app package (4.0K)
    /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:52:in
    `build_pack': Unable to detect a supported application type
    (RuntimeError)
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:26:in
    `block in compile_with_timeout'
    from /usr/lib/ruby/1.9.1/timeout.rb:68:in `timeout'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:25:in
    `compile_with_timeout'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:16:in
    `block in stage_application'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:12:in `chdir'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:12:in
    `stage_application'
    from /var/vcap/packages/dea_next/buildpacks/bin/run:10:in `<main>'
    Checking prova...
    Application failed to stage
  • Iwasaki Yudai at May 24, 2013 at 11:56 am
    It seems your application is not compatible with buildpack support.
    You may need to modify your application to push it to CFv2.

    The official docs would be helpful (but not completed yet) to make your
    application workable.

    * http://docs.cloudfoundry.com/docs/using/deploying-apps/index.html

    You can also consult Heroku's documentation. Heroku's build pack support is
    almost compatible with that of CF.

    I cannot figure out why your instance is unstable. If you re-installed your
    instance without re-installing your OS, it may be the reason.
    On Friday, May 24, 2013 7:24:17 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/24 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>:
    Try 'cf info --all'. It shows other information on your CF instance.
    If you cannot push any applications, let me know.
    $ cf info --all
    Getting services... OK

    Cloud Foundry sponsored by Pivotal

    target: http://api.vcap.me
    version: 2
    support: http://support.cloudfoundry.com

    user: mi...@vcap.me <javascript:>

    service version provider plans description
    mysql 5.5 core 100 MySQL database
    postgresql 9.1 core 100 PostgreSQL database (vFabric)


    $ cf restart prova
    Using manifest file manifest.yml

    Stopping prova... OK

    Starting prova... OK
    -----> Downloaded app package (4.0K)
    /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:52:in
    `build_pack': Unable to detect a supported application type
    (RuntimeError)
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:26:in
    `block in compile_with_timeout'
    from /usr/lib/ruby/1.9.1/timeout.rb:68:in `timeout'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:25:in
    `compile_with_timeout'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:16:in
    `block in stage_application'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:12:in `chdir'
    from /var/vcap/packages/dea_next/buildpacks/lib/buildpack.rb:12:in
    `stage_application'
    from /var/vcap/packages/dea_next/buildpacks/bin/run:10:in `<main>'
    Checking prova...
    Application failed to stage
  • Gandalf Corvotempesta at May 24, 2013 at 12:15 pm

    2013/5/24 Iwasaki Yudai <iwasaki.yudai@gmail.com>:
    It seems your application is not compatible with buildpack support.
    You may need to modify your application to push it to CFv2.
    I'm trying to publish a plain HTML webpage.
    Do I need a buildpack also for plain-html ?

    Do you also know if a buildpack for PHP is available ?
  • Gandalf Corvotempesta at May 24, 2013 at 12:17 pm

    2013/5/24 Gandalf Corvotempesta <gandalf.corvotempesta@gmail.com>:
    Do you also know if a buildpack for PHP is available ?
    Solved with this:
    https://github.com/heroku/heroku-buildpack-php
  • Gandalf Corvotempesta at May 28, 2013 at 10:33 pm
    Il giorno 11/mag/2013 16:11, "Iwasaki Yudai" <iwasaki.yudai@gmail.com> ha
    scritto:
    I'll write some chef specs or scripts to install all components on a
    single server with a single command.

    Did you have time to create some chef recipes?
    With some chef recipes that will install a whole infrastructure on multiple
    server, I think that nise bosh could became the standard for installing CF.

    It will be nice to have some recipes that allows, for example, to install
    20 deas with a single command in an existing chef infrastructure.
  • Iwasaki Yudai at May 29, 2013 at 1:45 am
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially, test
    them.
    On Wednesday, May 29, 2013 7:33:44 AM UTC+9, Gandalf Corvotempesta wrote:


    Il giorno 11/mag/2013 16:11, "Iwasaki Yudai" <iwasak...@gmail.com<javascript:>>
    ha scritto:
    I'll write some chef specs or scripts to install all components on a
    single server with a single command.

    Did you have time to create some chef recipes?
    With some chef recipes that will install a whole infrastructure on
    multiple server, I think that nise bosh could became the standard for
    installing CF.

    It will be nice to have some recipes that allows, for example, to install
    20 deas with a single command in an existing chef infrastructure.
  • Gandalf Corvotempesta at May 29, 2013 at 9:14 am
    2013/5/29 Iwasaki Yudai <iwasaki.yudai@gmail.com>
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially, test
    them.
    Ok, i'll try to create some automation tasks on my own.
    Just some clarification bacause i'm not a bosh user:

    These two lines do the magic, installing the whole system.
    The first line will install a standard "micro" template and the second one
    it will also install a "micro_ng" template:

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../micro.yml micro
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh --keep-monit-files -y
    ../cf-release ../micro.yml micro_ng


    In manifest we have:

    jobs:
    - name: micro
       template:
       - postgres
       - nats
       - dea_next
       - health_manager_next
       - serialization_data_server
       - uaa
       - vcap_redis
       - mysql_node_ng
       - mysql_gateway
       - postgresql_node_ng
       - postgresql_gateway
    - name: micro_ng
       template:
       - cloud_controller_ng
       - gorouter


    this will define both jobs with all templates needed by each one, right?
    So, calling "micro" will install:
    postgres, nats, dea_next and so on.

    But where these are defined ?
    In manifest I can see:

    properties:
        micro: true


    What's this?

    I can also see:

    dea_next:
         memory_mb: 2048
         disk_mb: 8192
         num_instances: 16

    is this the only thing needed to install a dea_next?

    What If I would like to create a job for installing a standalone MySQL Node?
    Adding something like this to the manifest (inside 'jobs') should be enough?

      - name: mysql_node
        template:
        - nats
        - mysql_node_ng
        - mysql_gateway

    and then call

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../full_manifest.yml mysql_node

    But I don't see any references to "mysql_node_ng" but only to "mysql_node".
  • Iwasaki Yudai at May 29, 2013 at 1:35 pm
    Your understanding is correct. You can add a mysql job using the
    definition you wrote.

    Each item in the list of the jobs entry is called "Job template" and it
    simply corresponds to a directory in the "jobs" directory of your
    cf-release repository. You can find the directories in the jobs directory
    which have the same names in the list.

    The definition of the word "job" is little confusing now. Before the job
    collocation was introduced, 1VM could have had only 1 "job" that defined in
    the "jobs" directory.
    However, the current meaning of the "job" is a set of "job templates". You
    can now deploy multiple "job templates" (former "jobs") to a single VM.
    However the directory name is still "jobs" and this makes confusion. Job
    templates are sometimes called "jobs" and sometimes called "(job)
    templates". Additionally, job templates have template files in their
    directory, so you also need to be careful when you see the word "templates".

    The values in the "properties" entry are used to fill the variables in
    template files. You can find many variables in template files of job
    templates.

    On Wednesday, May 29, 2013 6:14:03 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/29 Iwasaki Yudai <iwasak...@gmail.com <javascript:>>
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially, test
    them.
    Ok, i'll try to create some automation tasks on my own.
    Just some clarification bacause i'm not a bosh user:

    These two lines do the magic, installing the whole system.
    The first line will install a standard "micro" template and the second one
    it will also install a "micro_ng" template:


    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release ../micro.yml micro
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh --keep-monit-files -y ../cf-release ../micro.yml micro_ng


    In manifest we have:

    jobs:
    - name: micro
    template:
    - postgres

    - nats
    - dea_next
    - health_manager_next
    - serialization_data_server
    - uaa

    - vcap_redis
    - mysql_node_ng
    - mysql_gateway
    - postgresql_node_ng
    - postgresql_gateway
    - name: micro_ng
    template:
    - cloud_controller_ng
    - gorouter


    this will define both jobs with all templates needed by each one, right?
    So, calling "micro" will install:
    postgres, nats, dea_next and so on.

    But where these are defined ?
    In manifest I can see:

    properties:
    micro: true


    What's this?

    I can also see:

    dea_next:
    memory_mb: 2048
    disk_mb: 8192
    num_instances: 16

    is this the only thing needed to install a dea_next?

    What If I would like to create a job for installing a standalone MySQL Node?
    Adding something like this to the manifest (inside 'jobs') should be enough?

    - name: mysql_node
    template:
    - nats
    - mysql_node_ng
    - mysql_gateway

    and then call

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../full_manifest.yml mysql_node

    But I don't see any references to "mysql_node_ng" but only to "
    mysql_node".







  • Gandalf Corvotempesta at May 29, 2013 at 2:35 pm
    2013/5/29 Iwasaki Yudai <iwasaki.yudai@gmail.com>
    Your understanding is correct. You can add a mysql job using the
    definition you wrote.

    Each item in the list of the jobs entry is called "Job template" and it
    simply corresponds to a directory in the "jobs" directory of your
    cf-release repository. You can find the directories in the jobs directory
    which have the same names in the list. <cut>
    The values in the "properties" entry are used to fill the variables in
    template files. You can find many variables in template files of job
    templates.
    Ok, now it's clear.
  • Iwasaki Yudai at May 29, 2013 at 4:56 pm
    I've created a new repository with simpler installer scripts.

    * https://github.com/yudai/cf_nise_installer

    Now you can install an CF instance to both your local machine and a vagrant
    VM.
    On Wednesday, May 29, 2013 10:35:26 PM UTC+9, Iwasaki Yudai wrote:

    Your understanding is correct. You can add a mysql job using the
    definition you wrote.

    Each item in the list of the jobs entry is called "Job template" and it
    simply corresponds to a directory in the "jobs" directory of your
    cf-release repository. You can find the directories in the jobs directory
    which have the same names in the list.

    The definition of the word "job" is little confusing now. Before the job
    collocation was introduced, 1VM could have had only 1 "job" that defined
    in the "jobs" directory.
    However, the current meaning of the "job" is a set of "job templates". You
    can now deploy multiple "job templates" (former "jobs") to a single VM.
    However the directory name is still "jobs" and this makes confusion. Job
    templates are sometimes called "jobs" and sometimes called "(job)
    templates". Additionally, job templates have template files in their
    directory, so you also need to be careful when you see the word "templates".

    The values in the "properties" entry are used to fill the variables in
    template files. You can find many variables in template files of job
    templates.

    On Wednesday, May 29, 2013 6:14:03 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/29 Iwasaki Yudai <iwasak...@gmail.com>
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially, test
    them.
    Ok, i'll try to create some automation tasks on my own.
    Just some clarification bacause i'm not a bosh user:

    These two lines do the magic, installing the whole system.
    The first line will install a standard "micro" template and the second
    one it will also install a "micro_ng" template:


    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release ../micro.yml micro
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh --keep-monit-files -y ../cf-release ../micro.yml micro_ng


    In manifest we have:

    jobs:
    - name: micro
    template:
    - postgres

    - nats
    - dea_next
    - health_manager_next
    - serialization_data_server
    - uaa

    - vcap_redis
    - mysql_node_ng
    - mysql_gateway
    - postgresql_node_ng
    - postgresql_gateway
    - name: micro_ng
    template:
    - cloud_controller_ng
    - gorouter


    this will define both jobs with all templates needed by each one, right?
    So, calling "micro" will install:
    postgres, nats, dea_next and so on.

    But where these are defined ?
    In manifest I can see:

    properties:
    micro: true


    What's this?

    I can also see:

    dea_next:
    memory_mb: 2048
    disk_mb: 8192
    num_instances: 16

    is this the only thing needed to install a dea_next?

    What If I would like to create a job for installing a standalone MySQL Node?
    Adding something like this to the manifest (inside 'jobs') should be enough?

    - name: mysql_node
    template:
    - nats
    - mysql_node_ng
    - mysql_gateway

    and then call

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../full_manifest.yml mysql_node

    But I don't see any references to "mysql_node_ng" but only to "
    mysql_node".







  • Mgarciap at May 30, 2013 at 3:45 am
    Hi Iwasaki,

    I tried it out on a ubuntu 12.04.4 server and got this while running the
    packaging script for ruby:
    #[1794] INFO: Running the packaging script for ruby

    #[1794] INFO: An error occurred while compiling ruby
    #[1794] INFO: + set -e
    + tar xzf ruby/ruby-1.9.2-p180.tar.gz
    + cd ruby-1.9.2-p180
    ...
    ...
    compiling openssl
    make[1]: Entering directory
    `/var/vcap/data/compile/ruby/ruby-1.9.2-p180/ext/openssl'
    gcc -I. -I../../.ext/include/x86_64-linux -I../.././include
    -I../.././ext/openssl -DRUBY_EXTCONF_H=\"extconf.h\"
    -I/var/vcap/packages/libyaml/include -fPIC -O3 -ggdb -Wextra
    -Wno-unused-parameter -Wno-parentheses -Wpointer-arith -Wwrite-strings
    -Wno-missing-field-initializers -Wno-long-long -o ossl_asn1.o -c
    ossl_asn1.c
    gcc -I. -I../../.ext/include/x86_64-linux -I../.././include
    -I../.././ext/openssl -DRUBY_EXTCONF_H=\"extconf.h\"
    -I/var/vcap/packages/libyaml/include -fPIC -O3 -ggdb -Wextra
    -Wno-unused-parameter -Wno-parentheses -Wpointer-arith -Wwrite-strings
    -Wno-missing-field-initializers -Wno-long-long -o ossl_ssl.o -c ossl_ssl.c
    ossl_ssl.c:110:1: error: ‘SSLv2_method’ undeclared here (not in a function)
    ossl_ssl.c:111:1: error: ‘SSLv2_server_method’ undeclared here (not in a
    function)
    ossl_ssl.c:112:1: error: ‘SSLv2_client_method’ undeclared here (not in a
    function)
    make[1]: *** [ossl_ssl.o] Error 1
    make[1]: Leaving directory
    `/var/vcap/data/compile/ruby/ruby-1.9.2-p180/ext/openssl'
    make: *** [mkmain.sh] Error 1


    As far as I read, openssl does not longer support ssl V2 by default. If if
    rebuild and reinstall the openssl package removing the 'no-ssl2...' args
    and settings it should work.

    Is there any other solution?
    thanks
    On Wednesday, May 29, 2013 1:56:40 PM UTC-3, Iwasaki Yudai wrote:

    I've created a new repository with simpler installer scripts.

    * https://github.com/yudai/cf_nise_installer

    Now you can install an CF instance to both your local machine and a
    vagrant VM.
    On Wednesday, May 29, 2013 10:35:26 PM UTC+9, Iwasaki Yudai wrote:

    Your understanding is correct. You can add a mysql job using the
    definition you wrote.

    Each item in the list of the jobs entry is called "Job template" and it
    simply corresponds to a directory in the "jobs" directory of your
    cf-release repository. You can find the directories in the jobs directory
    which have the same names in the list.

    The definition of the word "job" is little confusing now. Before the job
    collocation was introduced, 1VM could have had only 1 "job" that defined
    in the "jobs" directory.
    However, the current meaning of the "job" is a set of "job templates".
    You can now deploy multiple "job templates" (former "jobs") to a single VM.
    However the directory name is still "jobs" and this makes confusion. Job
    templates are sometimes called "jobs" and sometimes called "(job)
    templates". Additionally, job templates have template files in their
    directory, so you also need to be careful when you see the word "templates".

    The values in the "properties" entry are used to fill the variables in
    template files. You can find many variables in template files of job
    templates.

    On Wednesday, May 29, 2013 6:14:03 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/29 Iwasaki Yudai <iwasak...@gmail.com>
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially, test
    them.
    Ok, i'll try to create some automation tasks on my own.
    Just some clarification bacause i'm not a bosh user:

    These two lines do the magic, installing the whole system.
    The first line will install a standard "micro" template and the second
    one it will also install a "micro_ng" template:


    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release ../micro.yml micro
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh --keep-monit-files -y ../cf-release ../micro.yml micro_ng


    In manifest we have:

    jobs:
    - name: micro
    template:
    - postgres

    - nats
    - dea_next
    - health_manager_next
    - serialization_data_server
    - uaa

    - vcap_redis
    - mysql_node_ng
    - mysql_gateway
    - postgresql_node_ng
    - postgresql_gateway
    - name: micro_ng
    template:
    - cloud_controller_ng
    - gorouter


    this will define both jobs with all templates needed by each one, right?
    So, calling "micro" will install:
    postgres, nats, dea_next and so on.

    But where these are defined ?
    In manifest I can see:

    properties:
    micro: true


    What's this?

    I can also see:

    dea_next:
    memory_mb: 2048
    disk_mb: 8192
    num_instances: 16

    is this the only thing needed to install a dea_next?

    What If I would like to create a job for installing a standalone MySQL Node?
    Adding something like this to the manifest (inside 'jobs') should be enough?

    - name: mysql_node
    template:
    - nats
    - mysql_node_ng
    - mysql_gateway

    and then call

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../full_manifest.yml mysql_node

    But I don't see any references to "mysql_node_ng" but only to "
    mysql_node".







  • Iwasaki Yudai at May 30, 2013 at 3:58 am
    cf-release is runnable only on Lucid64.
    This is not the limitation of the installer script, so I can do nothing.
    Please use 10.04 instead of 12.04.

    If you want to use 12.04 for some reason, you need to fork cf-release and
    add some tricks to your packaging file of the ruby package.
    Add smething like this before the ./configure command in
    packages/ruby/packaging:
    sed -i 's/\(OSSL_SSL_METHOD_ENTRY(SSLv2[^3]\)/\/\/\1/g'
    ./ext/openssl/ossl_ssl.c

    And then, modify scripts/clone_cf_release.sh of cf_nise_installer to point
    your own cf-release repository.
    On Thursday, May 30, 2013 12:45:12 PM UTC+9, mgar...@gmail.com wrote:

    Hi Iwasaki,

    I tried it out on a ubuntu 12.04.4 server and got this while running the
    packaging script for ruby:
    #[1794] INFO: Running the packaging script for ruby

    #[1794] INFO: An error occurred while compiling ruby
    #[1794] INFO: + set -e
    + tar xzf ruby/ruby-1.9.2-p180.tar.gz
    + cd ruby-1.9.2-p180
    ...
    ...
    compiling openssl
    make[1]: Entering directory
    `/var/vcap/data/compile/ruby/ruby-1.9.2-p180/ext/openssl'
    gcc -I. -I../../.ext/include/x86_64-linux -I../.././include
    -I../.././ext/openssl -DRUBY_EXTCONF_H=\"extconf.h\"
    -I/var/vcap/packages/libyaml/include -fPIC -O3 -ggdb -Wextra
    -Wno-unused-parameter -Wno-parentheses -Wpointer-arith -Wwrite-strings
    -Wno-missing-field-initializers -Wno-long-long -o ossl_asn1.o -c
    ossl_asn1.c
    gcc -I. -I../../.ext/include/x86_64-linux -I../.././include
    -I../.././ext/openssl -DRUBY_EXTCONF_H=\"extconf.h\"
    -I/var/vcap/packages/libyaml/include -fPIC -O3 -ggdb -Wextra
    -Wno-unused-parameter -Wno-parentheses -Wpointer-arith -Wwrite-strings
    -Wno-missing-field-initializers -Wno-long-long -o ossl_ssl.o -c ossl_ssl.c
    ossl_ssl.c:110:1: error: ‘SSLv2_method’ undeclared here (not in a function)
    ossl_ssl.c:111:1: error: ‘SSLv2_server_method’ undeclared here (not in a
    function)
    ossl_ssl.c:112:1: error: ‘SSLv2_client_method’ undeclared here (not in a
    function)
    make[1]: *** [ossl_ssl.o] Error 1
    make[1]: Leaving directory
    `/var/vcap/data/compile/ruby/ruby-1.9.2-p180/ext/openssl'
    make: *** [mkmain.sh] Error 1


    As far as I read, openssl does not longer support ssl V2 by default. If if
    rebuild and reinstall the openssl package removing the 'no-ssl2...' args
    and settings it should work.

    Is there any other solution?
    thanks
    On Wednesday, May 29, 2013 1:56:40 PM UTC-3, Iwasaki Yudai wrote:

    I've created a new repository with simpler installer scripts.

    * https://github.com/yudai/cf_nise_installer

    Now you can install an CF instance to both your local machine and a
    vagrant VM.
    On Wednesday, May 29, 2013 10:35:26 PM UTC+9, Iwasaki Yudai wrote:

    Your understanding is correct. You can add a mysql job using the
    definition you wrote.

    Each item in the list of the jobs entry is called "Job template" and it
    simply corresponds to a directory in the "jobs" directory of your
    cf-release repository. You can find the directories in the jobs directory
    which have the same names in the list.

    The definition of the word "job" is little confusing now. Before the
    job collocation was introduced, 1VM could have had only 1 "job" that
    defined in the "jobs" directory.
    However, the current meaning of the "job" is a set of "job templates".
    You can now deploy multiple "job templates" (former "jobs") to a single VM.
    However the directory name is still "jobs" and this makes confusion. Job
    templates are sometimes called "jobs" and sometimes called "(job)
    templates". Additionally, job templates have template files in their
    directory, so you also need to be careful when you see the word "templates".

    The values in the "properties" entry are used to fill the variables in
    template files. You can find many variables in template files of job
    templates.

    On Wednesday, May 29, 2013 6:14:03 PM UTC+9, Gandalf Corvotempesta wrote:

    2013/5/29 Iwasaki Yudai <iwasak...@gmail.com>
    I'm sorry but I myself have no plan to create chef specs.
    I currently do not have enough resource to create and, especially,
    test them.
    Ok, i'll try to create some automation tasks on my own.
    Just some clarification bacause i'm not a bosh user:

    These two lines do the magic, installing the whole system.
    The first line will install a standard "micro" template and the second
    one it will also install a "micro_ng" template:


    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release ../micro.yml micro
    sudo env PATH=$PATH bundle exec ./bin/nise-bosh --keep-monit-files -y ../cf-release ../micro.yml micro_ng


    In manifest we have:

    jobs:
    - name: micro
    template:
    - postgres

    - nats
    - dea_next
    - health_manager_next
    - serialization_data_server
    - uaa

    - vcap_redis
    - mysql_node_ng
    - mysql_gateway
    - postgresql_node_ng
    - postgresql_gateway
    - name: micro_ng
    template:
    - cloud_controller_ng
    - gorouter


    this will define both jobs with all templates needed by each one, right?
    So, calling "micro" will install:
    postgres, nats, dea_next and so on.

    But where these are defined ?
    In manifest I can see:

    properties:
    micro: true


    What's this?

    I can also see:

    dea_next:
    memory_mb: 2048
    disk_mb: 8192
    num_instances: 16

    is this the only thing needed to install a dea_next?

    What If I would like to create a job for installing a standalone MySQL Node?
    Adding something like this to the manifest (inside 'jobs') should be enough?

    - name: mysql_node
    template:
    - nats
    - mysql_node_ng
    - mysql_gateway

    and then call

    sudo env PATH=$PATH bundle exec ./bin/nise-bosh -y ../cf-release
    ../full_manifest.yml mysql_node

    But I don't see any references to "mysql_node_ng" but only to "
    mysql_node".







Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupvcap-dev @
postedMar 11, '13 at 8:51a
activeJun 1, '13 at 1:34a
posts59
users10

People

Translate

site design / logo © 2021 Grokbase