From d0b167bd075eda5ffa104229103dfad772e9f403 Mon Sep 17 00:00:00 2001 From: Chengcheng Mu Date: Tue, 18 Aug 2015 10:46:23 +0200 Subject: fixed some issues to boot VM on GCE corrected openshift master config, commented infra correct list and terminate, it was bugged in case where no instance were terminated Using openshift-sdn for gce new join_node playbook for gce openstack/hosts/nova.py is now taking the nova/ini of its directory and not the directory of execution of bin/cluster add fix of ICMP reject rules Avoid a recursive loop Jenkins image was renamed Default masters to t2.medium instead of t2.small Fix a minor bug involving AWS ENV Keys * If a user forgot to set their AWS keys, we'd get a non descriptive error about a variable not being set * This patch uses the correct variable so the error message is more informative delete some fix that are needed anymore (selinux, iptables rules for sdn) GCE : all variables needed are in gce.ini, it will be used by bin/cluster (now check better the presence of gce.init in the default place or use GCE_INI_PATH to locate it ), also by gce.ini openshift_node_labels : get from oo_option fix syntax error in bin/cluster fix lookup for openshift_node_labels Adding desc, multiplier, and units to zabbix item Adding capability to have descriptions on triggers updated triggers and items to have better descriptions and multipliers Move openshift_data_dir to a fact based on deployment_type Previously this was being set to /var/lib/origin regardless of deployment_type which isn't correct given that existing 'enterprise' and 'online' deployments would have been deployed with /var/lib/openshift Verify again that ansible version is different than 1.9.0 and 1.9.0.1 bin/cluste does not take -a and -s anymore fix master_public_api_url : using by default a correct url Really fixed master public api url this time Really fixed master public api url this time uncommented infra deployment like before fixed again masterpublicurl in a template README_GCE.md : use GCE_INI_PATH in order to locate gce.ini, update description of gce.ini --- README_GCE.md | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) (limited to 'README_GCE.md') diff --git a/README_GCE.md b/README_GCE.md index f6c5138c1..50f8ade70 100644 --- a/README_GCE.md +++ b/README_GCE.md @@ -39,6 +39,13 @@ Create a gce.ini file for GCE * gce_service_account_pem_file_path - Full path from previous steps * gce_project_id - Found in "Projects", it list all the gce projects you are associated with. The page lists their "Project Name" and "Project ID". You want the "Project ID" +Mandatory customization variables (check the values according to your tenant): +* zone = europe-west1-d +* network = default +* gce_machine_type = n1-standard-2 +* gce_machine_image = preinstalled-slave-50g-v5 + + 1. vi ~/.gce/gce.ini 1. make the contents look like this: ``` @@ -46,11 +53,15 @@ Create a gce.ini file for GCE gce_service_account_email_address = long...@developer.gserviceaccount.com gce_service_account_pem_file_path = /full/path/to/project_id-gce_key_hash.pem gce_project_id = project_id +zone = europe-west1-d +network = default +gce_machine_type = n1-standard-2 +gce_machine_image = preinstalled-slave-50g-v5 + ``` -1. Setup a sym link so that gce.py will pick it up (link must be in same dir as gce.py) +1. Define the environment variable GCE_INI_PATH so gce.py can pick it up and bin/cluster can also read it ``` - cd openshift-ansible/inventory/gce - ln -s ~/.gce/gce.ini gce.ini +export GCE_INI_PATH=~/.gce/gce.ini ``` -- cgit v1.2.3