Bosh & Cf release & Properties

[gwenn.e...@rakops.com]

Hi, 

I manage to deploy CF V2 using bosh behind a firewall and using a proxy.

So in my bosh manifest I put the env property

properties:

  env:

    http_proxy: "http://dev-proxy.db.rakuten.co.jp:9501"

    https_proxy: "http://dev-proxy.db.rakuten.co.jp:9501"

    no_proxy: "127.0.0.1"

But it's seems this is not in the final ctl of the CC.

CTL : https://github.com/cloudfoundry/cloud_controller_ng/blob/master/bosh-templates/cloud_controller_api_ctl.erb

Do you have any idea why ?

Thanks

[dmi...@pivotallabs.com]

Please provide some additional details:

- Did you deployment successfully finish?

- Which cf-release version are you using?

- Which BOSH director and stemcell versions are you using?

- Is `properties.env` section scoped globally or under a specific job in your deployment manifest?

- Are there any values that do get successfully interpolated in the final ctl script?

- Since CC ctl script does not contain any sensitive info could you attach it below?

Thank you,

-dk

[gwenn.e...@rakops.com]

hi thanks for your reply,

- Did you deployment successfully finish?

==> Yes the deployment is succesfull done.

- Which cf-release version are you using?

==> 173

- Which BOSH director and stemcell versions are you using?

==> Director :  Version    1.2611.0 (00000000)

==> Stemcell: bosh-stemcell-2611-vsphere-esxi-ubuntu-lucid-go_agent

- Is `properties.env` section scoped globally or under a specific job in your deployment manifest?

==> Yes is globaly

- Are there any values that do get successfully interpolated in the final ctl script?

==> Yes all others values

- Since CC ctl script does not contain any sensitive info could you attach it below?

==> Ok let me some time I need redeploy :/

Thanks you gwenn

[gwenn.e...@rakops.com]

Here my cloudcontroller ctl

#!/bin/bash

export CC_JOB_DIR=/var/vcap/jobs/cloud_controller_ng

export PATH=/var/vcap/packages/ruby/bin:$PATH

CC_PACKAGE_DIR=/var/vcap/packages/cloud_controller_ng

RUN_DIR=/var/vcap/sys/run/cloud_controller_ng

LOG_DIR=/var/vcap/sys/log/cloud_controller_ng

PIDFILE=$RUN_DIR/cloud_controller_ng.pid

export CONFIG_DIR=$CC_JOB_DIR/config

export CLOUD_CONTROLLER_NG_CONFIG=$CONFIG_DIR/cloud_controller_ng.yml

export BUNDLE_GEMFILE=$CC_PACKAGE_DIR/cloud_controller_ng/Gemfile

export HOME=/home/vcap # rake needs it to be set to run tasks

export TMPDIR=/var/vcap/data/cloud_controller_ng/tmp

export C_INCLUDE_PATH=/var/vcap/packages/libpq/include:$C_INCLUDE_PATH

export LIBRARY_PATH=/var/vcap/packages/libpq/lib:$LIBRARY_PATH

export LANG=en_US.UTF-8

source /var/vcap/packages/common/utils.sh

case $1 in

  start)

    pid_guard $PIDFILE "Cloud controller ng"

    mkdir -p $RUN_DIR

    mkdir -p $LOG_DIR

    mkdir -p $TMPDIR

    chown vcap:vcap $RUN_DIR

    chown vcap:vcap $LOG_DIR

    chown vcap:vcap $TMPDIR

    source $CC_JOB_DIR/bin/handle_nfs_or_local_blobstore.sh

    # Configure the core file location

    mkdir -p /var/vcap/sys/cores

    chown vcap:vcap /var/vcap/sys/cores

    echo /var/vcap/sys/cores/core-%e-%s-%p-%t > /proc/sys/kernel/core_pattern

    ulimit -c unlimited

    /var/vcap/packages/syslog_aggregator/setup_syslog_forwarder.sh $CONFIG_DIR

    # Run the migrations only on the first CC launch

      cd $CC_PACKAGE_DIR/cloud_controller_ng

      chpst -u vcap:vcap bundle exec rake db:migrate

      if [ $? != 0 ]; then

        echo "Migrations failed"

        exit 1

      fi

      # TODO: remove the -m when we have a stand alone utility for loading

      # frameworks etc, and call those utilities here.  See corresponding

      # comments in cc runner.

      exec chpst -u vcap:vcap $CC_PACKAGE_DIR/cloud_controller_ng/bin/cloud_controller -m \

        -c $CLOUD_CONTROLLER_NG_CONFIG

    ;;

  stop)

    kill_and_wait $PIDFILE

    ;;

  *)

    echo "Usage: cloud_controller_ng_ctl {start|stop}"

    ;;

esac

On Tuesday, July 8, 2014 2:37:51 PM UTC+9, dmi...@pivotallabs.com wrote: