network validation to ping test each interface

This patches wires in a new "all nodes" validation resource
that can be used to add validations that occur early on
during the deployment process. This occurs after the nodes
have been brought online and the initial networks
have been configured but before any "post" (puppet, etc.)
sort of configuration has been executed.

A initial validation script has been added to ping test network IPs
on each network. When using network isolation this will ensure
network connectivity (vlans, etc) are working on each
node and if not the heat stack will fail early, allowing
time to fix the network connections and retry the
stack creation via an update.

Change-Id: I63cf95b27e8ad2aed48718cf84df5f324780e597
Co-Authored-By: Ian Main <imain@redhat.com>
Co-Authored-By: Ryan Hallisey <rhallise@redhat.com>

5 files changed, 110 insertions, 0 deletions

diff --git a/all-nodes-validation.yaml b/all-nodes-validation.yaml
new file mode 100644
index 00000000..a7383375
--- /dev/null
+++ b/all-nodes-validation.yaml
@@ -0,0 +1,27 @@
+heat_template_version: 2015-04-30
+
+description: >
+  Software Config to drive validations that occur on all nodes.
+  Note, you need the heat-config-script element built into your
+  images, due to the script group below.
+
+parameters:
+  PingTestIps:
+    default: ''
+    description: A string containing a space separated list of IP addresses used to ping test each available network interface.
+    type: string
+
+resources:
+  AllNodesValidationsImpl:
+    type: OS::Heat::SoftwareConfig
+    properties:
+      group: script
+      inputs:
+        - name: ping_test_ips
+          default: {get_param: PingTestIps}
+      config: {get_file: ./validation-scripts/all-nodes.sh}
+
+outputs:
+  OS::stack_id:
+    description: The ID of the AllNodesValidationsImpl resource.
+    value: {get_resource: AllNodesValidationsImpl}
diff --git a/overcloud-resource-registry-puppet.yaml b/overcloud-resource-registry-puppet.yaml
index 17c1470f..91b91ced 100644
--- a/overcloud-resource-registry-puppet.yaml
+++ b/overcloud-resource-registry-puppet.yaml
@@ -74,5 +74,8 @@ resource_registry:
   # Port assignments for service virtual IPs for the controller role
   OS::TripleO::Controller::Ports::RedisVipPort: network/ports/ctlplane_vip.yaml
 
+  # validation resources
+  OS::TripleO::AllNodes::Validation: all-nodes-validation.yaml
+
 parameter_defaults:
   EnablePackageInstall: false
diff --git a/overcloud-resource-registry.yaml b/overcloud-resource-registry.yaml
index 78607b51..76e0c431 100644
--- a/overcloud-resource-registry.yaml
+++ b/overcloud-resource-registry.yaml
@@ -64,3 +64,6 @@ resource_registry:
 
   # Port assignments for service virtual IPs for the controller role
   OS::TripleO::Controller::Ports::RedisVipPort: network/ports/noop.yaml
+
+  # validation resources
+  OS::TripleO::AllNodes::Validation: all-nodes-validation.yaml
diff --git a/overcloud-without-mergepy.yaml b/overcloud-without-mergepy.yaml
index fd31c54d..193c7909 100644
--- a/overcloud-without-mergepy.yaml
+++ b/overcloud-without-mergepy.yaml
@@ -1226,6 +1226,54 @@ resources:
       config: {get_attr: [allNodesConfig, config_id]}
       servers: {get_attr: [CephStorage, attributes, nova_server_resource]}
 
+  # All Nodes Validations
+  AllNodesValidationConfig:
+    type: OS::TripleO::AllNodes::Validation
+    properties:
+      PingTestIps:
+        list_join:
+        - ' '
+        - - {get_attr: [Controller, resource.0.external_ip_address]}
+          - {get_attr: [Controller, resource.0.internal_api_ip_address]}
+          - {get_attr: [Controller, resource.0.storage_ip_address]}
+          - {get_attr: [Controller, resource.0.storage_mgmt_ip_address]}
+          - {get_attr: [Controller, resource.0.tenant_ip_address]}
+
+  ControllerAllNodesValidationDeployment:
+    type: OS::Heat::StructuredDeployments
+    depends_on: ControllerAllNodesDeployment
+    properties:
+      config: {get_resource: AllNodesValidationConfig}
+      servers: {get_attr: [Controller, attributes, nova_server_resource]}
+
+  ComputeAllNodesValidationDeployment:
+    type: OS::Heat::StructuredDeployments
+    depends_on: ComputeAllNodesDeployment
+    properties:
+      config: {get_resource: AllNodesValidationConfig}
+      servers: {get_attr: [Compute, attributes, nova_server_resource]}
+
+  BlockStorageAllNodesValidationDeployment:
+    type: OS::Heat::StructuredDeployments
+    depends_on: BlockStorageAllNodesDeployment
+    properties:
+      config: {get_resource: AllNodesValidationConfig}
+      servers: {get_attr: [BlockStorage, attributes, nova_server_resource]}
+
+  ObjectStorageAllNodesValidationDeployment:
+    type: OS::Heat::StructuredDeployments
+    depends_on: ObjectStorageAllNodesDeployment
+    properties:
+      config: {get_resource: AllNodesValidationConfig}
+      servers: {get_attr: [ObjectStorage, attributes, nova_server_resource]}
+
+  CephStorageAllNodesValidationDeployment:
+    type: OS::Heat::StructuredDeployments
+    depends_on: CephStorageAllNodesDeployment
+    properties:
+      config: {get_resource: AllNodesValidationConfig}
+      servers: {get_attr: [CephStorage, attributes, nova_server_resource]}
+
   # Nested stack deployment runs after all other controller deployments
   ControllerNodesPostDeployment:
     type: OS::TripleO::ControllerPostDeployment
diff --git a/validation-scripts/all-nodes.sh b/validation-scripts/all-nodes.sh
new file mode 100644
index 00000000..38a5a55e
--- /dev/null
+++ b/validation-scripts/all-nodes.sh
@@ -0,0 +1,29 @@
+#!/bin/bash
+
+# For each unique remote IP (specified via Heat) we check to
+# see if one of the locally configured networks matches and if so we
+# attempt a ping test on that networks remote IP.
+function ping_controller_ips() {
+  local REMOTE_IPS=$1
+
+  for REMOTE_IP in $(echo $REMOTE_IPS | sed -e "s| |\n|g" | sort -u); do
+
+    for LOCAL_NETWORK in $(ip r | grep -v default | cut -d " " -f 1); do
+       local LOCAL_CIDR=$(echo $LOCAL_NETWORK | cut -d "/" -f 2)
+       local LOCAL_NETMASK=$(ipcalc -m $LOCAL_NETWORK | grep NETMASK | cut -d "=" -f 2)
+       local REMOTE_NETWORK=$(ipcalc -np $REMOTE_IP $LOCAL_NETMASK | grep NETWORK | cut -d "=" -f 2)
+
+       if [ $REMOTE_NETWORK/$LOCAL_CIDR == $LOCAL_NETWORK ]; then
+         echo -n "Trying to ping $REMOTE_IP for local network $LOCAL_NETWORK..."
+         if ! ping -c 1 $REMOTE_IP &> /dev/null; then
+           echo "FAILURE"
+           echo "$REMOTE_IP is not pingable. Local Network: $LOCAL_NETWORK" >&2
+           exit 1
+         fi
+         echo "SUCCESS"
+       fi
+    done
+  done
+}
+
+ping_controller_ips "$ping_test_ips"