devtest_testenv

Note

This script is usually called from devtest.sh as devtest_testenv.sh $TE_DATAFILE so we should declare a JSONFILE variable (which equals to the first positional argument) explicitly.

JSONFILE=${JSONFILE:-$TE_DATAFILE}

1. Set HW resources for VMs used as ‘baremetal’ nodes. NODE_CPU is cpu count, NODE_MEM is memory (MB), NODE_DISK is disk size (GB), NODE_ARCH is architecture (i386, amd64). NODE_ARCH is used also for the seed VM. A note on memory sizing: TripleO images in raw form are currently ~2.7Gb, which means that a tight node will end up with a thrashing page cache during glance -> local + local -> raw operations. This significantly impairs performance. Of the four minimum VMs for TripleO simulation, two are nova baremetal nodes (seed and undercloud) and these need to be 3G or larger. The hypervisor host in the overcloud also needs to be a decent size or it cannot host more than one VM. The NODE_DISK is set to support building 5 overcloud nodes when not using Ironic. If you are building a larger overcloud than this without using Ironic you may need to increase NODE_DISK.

   NODE_CNT specifies how many VMs to define using virsh. NODE_CNT defaults to 15, or 0 if NODES_PATH is provided.

   32bit VMs

   NODE_CPU=1 NODE_MEM=3072 NODE_DISK=40 NODE_ARCH=i386
   

   For 64bit it is better to create VMs with more memory and storage because of increased memory footprint (we suggest 4GB):

   NODE_CPU=1 NODE_MEM=4096 NODE_DISK=40 NODE_ARCH=amd64
   

2. Configure a network for your test environment. This configures an openvswitch bridge and teaches libvirt about it.

   setup-network -n "$NUM" -b "$BRIDGE_NAMES"
   

3. Configure a seed VM. This VM has a disk image manually configured by later scripts, and hosts the statically configured seed which is used to bootstrap a full dynamically configured baremetal cloud. The seed VM specs can be configured with the environment variables SEED_CPU and SEED_MEM (MB). It defaults to the NODE_CPU and NODE_MEM values, since the seed is equivalent to an undercloud in resource requirements.

   NUMBERED_BRIDGE_NAMES=
   SEED_ARGS="-a $NODE_ARCH"
   if [ -n "$NUM" ]; then
       SEED_ARGS="$SEED_ARGS -o seed_${NUM}"
   fi
   if [ -n "$OVSBRIDGE" ]; then
       SEED_ARGS="$SEED_ARGS -p $OVSBRIDGE"
   fi
   for NAME in $BRIDGE_NAMES; do
       NUMBERED_BRIDGE_NAMES="$NUMBERED_BRIDGE_NAMES$NAME${NUM} "
   done
   # remove the last space
   NUMBERED_BRIDGE_NAMES=${NUMBERED_BRIDGE_NAMES% }
   
   SEED_CPU=${SEED_CPU:-${NODE_CPU}}
   SEED_MEM=${SEED_MEM:-${NODE_MEM}}
   

4. Clean up any prior environment. Unless the –keep-vms argument is passed to the script, VMs for the undercloud and overcloud are destroyed

   if [ -z "$KEEP_VMS" ]; then
       if [ -n "$NUM" ]; then
           cleanup-env -n $NUM -b "$BRIDGE_NAMES"
       else
           cleanup-env -b "$BRIDGE_NAMES"
       fi
   fi
   
   #Now start creating the new environment
   setup-seed-vm $SEED_ARGS -c ${SEED_CPU} -m $((1024 * ${SEED_MEM}))
   

5. What user will be used to ssh to run virt commands to control our emulated baremetal machines.

   SSH_USER=$(whoami)
   

6. What IP address to ssh to for virsh operations.

   HOSTIP=${HOSTIP:-192.168.122.1}
   

7. If a static SEEDIP is in use, define it here. If not defined it will be looked up in the ARP table by the seed MAC address during seed deployment.

   if [ -n "$NETS_PATH" ]; then
       # if the value is not set try the default 192.0.2.1.
       SEEDIP=$(jq '.["seed"]["ip"] // "192.0.2.1"' -r $NETS_PATH)
   else
       SEEDIP=${SEEDIP:-''}
   fi
   

8. Set the default bare metal power manager. By default devtest uses nova.virt.baremetal.virtual_power_driver.VirtualPowerManager to support a fully virtualized TripleO test environment. You may optionally customize this setting if you are using real baremetal hardware with the devtest scripts. This setting controls the power manager used in both the seed VM and undercloud for Nova Baremetal.

   POWER_MANAGER=${POWER_MANAGER:-'nova.virt.baremetal.virtual_power_driver.VirtualPowerManager'}
   

9. Ensure we can ssh into the host machine to turn VMs on and off. The private key we create will be embedded in the seed VM, and delivered dynamically by heat to the undercloud VM.

   # generate ssh authentication keys if they don't exist
   if [ ! -f $SSH_KEY ]; then
       ssh-keygen -t rsa -N "" -C virtual-power-key -f $SSH_KEY
   fi
   
   # make the local id_rsa_virt_power.pub be in ``.ssh/authorized_keys`` before
   # that is copied into images via ``local-config``
   if ! grep -qF "$(cat ${SSH_KEY}.pub)" ~/.ssh/authorized_keys; then
       cat ${SSH_KEY}.pub >> ~/.ssh/authorized_keys
       chmod 0600 ~/.ssh/authorized_keys
   fi
   

10. Wrap this all up into JSON.

    jq "." <<EOF > $JSONFILE
    {
        "arch":"$NODE_ARCH",
        "host-ip":"$HOSTIP",
        "power_manager":"$POWER_MANAGER",
        "seed-ip":"$SEEDIP",
        "ssh-key":"$(cat $SSH_KEY|sed 's,$,\\n,'|tr -d '\n')",
        "ssh-user":"$SSH_USER"
    }
    EOF
    

11. If you have an existing bare metal cloud network to use, use it. See baremetal-network section in Devtest test environment configuration for more details

    devtest_update_network.sh ${NETS_PATH:+--bm-networks $NETS_PATH} $JSONFILE
    

12. If you have an existing set of nodes to use, use them.

    if [ -n "$NODES_PATH" ]; then
        JSON=$(jq -s '.[0].nodes=.[1] | .[0]' $JSONFILE $NODES_PATH)
        echo "${JSON}" > $JSONFILE
    else
    

13. Create baremetal nodes for the test cluster. If the required number of VMs changes in future, you can run cleanup-env and then recreate with more nodes.

    create-nodes $NODE_CPU $NODE_MEM $NODE_DISK $NODE_ARCH $NODE_CNT $SSH_USER $HOSTIP $JSONFILE "$NUMBERED_BRIDGE_NAMES"