Skip to content

Commit

Permalink
Multi-cell adoption for remaining services
Browse files Browse the repository at this point in the history
Split edpm nodes into compute cells by 1:1 mapping it as
dataplane nodesets.

Use edpm_nodes var to describe compuptes for each cell,
instead of static host and ip vars that only used to work for
a single-cell standalone, or multi-node single cell cases.
Also explain EDPM net config requirements in vars.sample, when
it is used outside of ci-framework (local deployments).

Remove edpm_computes vars no longer used after moving stopping
control-plane tripleo services into edpm-ansible

Simplify ENV headers management by collecting in a single place.

Provide a variable to define the source cloud Ironic topology,
for any cells with Ironic services.

Align nova/libvirt and related services ordering in the
lists of services defined in multiple places, with those
specified in VA.

Align the names in the tests to follow the documented steps
to make the corresponding code easy discoverable.

Adjust storage/storageRequests values to make it better fitting
a multi-cell test scenarios. Also provide values in docs and
add a comment to adjust them as needed.

Stop ovn services only if active, or not missing (like on
the cell controllers)

Signed-off-by: Bohdan Dobrelia <[email protected]>
  • Loading branch information
bogdando committed Dec 10, 2024
1 parent 732bf83 commit 6a5c854
Show file tree
Hide file tree
Showing 21 changed files with 1,113 additions and 425 deletions.
137 changes: 131 additions & 6 deletions docs_dev/assemblies/development_environment.adoc
Original file line number Diff line number Diff line change
Expand Up @@ -187,6 +187,102 @@ https://openstack-k8s-operators.github.io/data-plane-adoption/dev/#_reset_the_en

'''

== Deploying TripleO With Multiple Cells

A TripleO Standalone setup creates only a single Nova v2 cell, with a combined controller and compute services on it.
In order to deploy multiple compute cells for adoption testing (without Ceph), create a 5 VMs, with the following requirements met:

* Named `edpm-compute-0` .. `edpm-compute-4`.
* Running RHEL 9.2, with RHOSP 17.1 repositories configured.
* Can login via SSH without a password as the root user, from the hypervisor host.
* User `zuul` is created, and can sudo without a password, and login via SSH without a password, from the hypervisor host.
* User `zuul` can login to `edpm-compute-1`, `edpm-compute-2`, `edpm-compute-3`, `edpm-compute-4` nodes via SSH without a password, from the `edpm-compute-0` node,
by using the generated `/home/zuul/.ssh/id_rsa` private key.
* RedHat registry credentials are exported on the hypervisor host.

Adjust the following commands for a repositories configuration tool of your choice:

[,bash]
----
export RH_REGISTRY_USER="<insert your registry.redhat.io user>"
export RH_REGISTRY_PWD="<insert your registry.redhat.io password>"
DEFAULT_CELL_NAME="cell3" <1>
RENAMED_CELLS="cell1 cell2 $DEFAULT_CELL_NAME"
cd ~/install_yamls/devsetup
cat <<EOF > /tmp/osp17_repos
# Use a tool of your choice:
# 1. Rhos-release example steps are only available from the internal RedHat network
# ... skipping download and install steps ...
# sudo rhos-release -x
# sudo rhos-release 17.1
# 2. Subscription-manager example steps require an active registration
# subscription-manager release --set=9.2
# subscription-manager repos --disable=*
# sudo subscription-manager repos \
# --enable=rhel-9-for-x86_64-baseos-eus-rpms \
# --enable=rhel-9-for-x86_64-appstream-eus-rpms \
# --enable=rhel-9-for-x86_64-highavailability-eus-rpms \
# --enable=openstack-17.1-for-rhel-9-x86_64-rpms \
# --enable=rhceph-6-tools-for-rhel-9-x86_64-rpms \
# --enable=fast-datapath-for-rhel-9-x86_64-rpms
# firstboot commands
sudo dnf install -y git curl wget podman python3-tripleoclient openvswitch3.1 NetworkManager-initscripts-updown \
sudo dnf install -y util-linux cephadm driverctl lvm2 jq nftables iptables-nft openstack-heat-agents \
os-net-config python3-libselinux python3-pyyaml rsync tmpwatch sysstat iproute-tc
sudo dnf install -y puppet-tripleo puppet-headless
sudo dnf install -y openstack-selinux
EOF
export CENTOS_9_STREAM_URL=<insert url to rhel-guest-image-9.2.x86_64.qcow2>
export NTP_SERVER=<insert ntp server of your choice>
export MANILA_ENABLED=false
export EDPM_COMPUTE_CEPH_ENABLED=false
export EDPM_COMPUTE_CEPH_NOVA=false
export EDPM_COMPUTE_CELLS=3
export STANDALONE_EXTRA_CMD="bash -c 'echo \"$RH_REGISTRY_PWD\" > ~/authfile; chmod 0600 ~/authfile; sudo dnf install -y podman; sudo /bin/podman login registry.redhat.io -u \"$RH_REGISTRY_USER\" --password-stdin < ~/authfile'"
export EDPM_FIRSTBOOT_EXTRA=/tmp/osp17_repos
export EDPM_TOTAL_NODES=1
export SKIP_TRIPLEO_REPOS=false
export EDPM_COMPUTE_NETWORK_IP=192.168.122.1
export HOST_PRIMARY_RESOLV_CONF_ENTRY=192.168.122.1
export BASE_DISK_FILENAME="rhel-9-base.qcow2"
EDPM_COMPUTE_SUFFIX=0 IP=192.168.122.100 EDPM_COMPUTE_DISK_SIZE=10 EDPM_COMPUTE_RAM=9 EDPM_COMPUTE_VCPUS=2 make edpm_compute
EDPM_COMPUTE_SUFFIX=1 IP=192.168.122.103 EDPM_COMPUTE_DISK_SIZE=17 EDPM_COMPUTE_RAM=12 EDPM_COMPUTE_VCPUS=4 make edpm_compute
EDPM_COMPUTE_SUFFIX=2 IP=192.168.122.106 EDPM_COMPUTE_DISK_SIZE=14 EDPM_COMPUTE_RAM=12 EDPM_COMPUTE_VCPUS=4 make edpm_compute
EDPM_COMPUTE_SUFFIX=3 IP=192.168.122.107 EDPM_COMPUTE_DISK_SIZE=12 EDPM_COMPUTE_RAM=4 EDPM_COMPUTE_VCPUS=2 make edpm_compute
EDPM_COMPUTE_SUFFIX=4 IP=192.168.122.109 EDPM_COMPUTE_DISK_SIZE=16 EDPM_COMPUTE_RAM=12 EDPM_COMPUTE_VCPUS=4 make edpm_compute
for n in 0 3 6 7 9; do
# w/a bad packages installation, if done by firstboot - resulting in rpm -V check failures in tripleo-ansible
ssh -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
[email protected]${n} dnf install -y openstack-selinux ';' \
dnf reinstall -y openstack-selinux
ssh -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
[email protected]${n} useradd --create-home --shell /bin/bash --groups root zuul ';' \
mkdir -p /home/zuul/.ssh
scp -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]${n}:/home/zuul/.ssh/id_rsa
ssh -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
[email protected]${n} cp /root/.ssh/authorized_keys /home/zuul/.ssh/authorized_keys
ssh -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
[email protected]${n} chown zuul: /home/zuul/.ssh/*
ssh -o StrictHostKeyChecking=false -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa \
[email protected]${n} echo "zuul ALL=NOPASSWD:ALL" '>' /etc/sudoers.d/zuul
done
make tripleo_deploy
for n in 0 1 2 3 4; do make standalone_snapshot EDPM_COMPUTE_SUFFIX=$n; done
----
<1> The source cloud default cell takes a new `$DEFAULT_CELL_NAME`. In a multi-cell adoption scenario, it may either retain its original name `default`, or become created as a last `cell<X>`.

== Network routing

Route VLAN20 to have access to the MariaDB cluster:
Expand Down Expand Up @@ -219,8 +315,10 @@ installing the package and copying the configuration file from the virtual machi

[,bash]
----
alias openstack="ssh -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected] OS_CLOUD=standalone openstack"
OS_CLOUD_NAME=standalone
alias openstack="ssh -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected] OS_CLOUD=$OS_CLOUD_NAME openstack"
----
For a multi-cell environment, set `OS_CLOUD_NAME` to `overcloud`.

=== Virtual machine steps

Expand Down Expand Up @@ -345,15 +443,28 @@ make openstack

== Performing the adoption procedure

To simplify the adoption procedure, copy the deployment passwords that
To simplify the adoption procedure with additional cells, copy and rename the deployment passwords that
you use in copy the deployment passwords that you use in the
https://openstack-k8s-operators.github.io/data-plane-adoption/user/#deploying-backend-services_migrating-databases[backend
services deployment phase of the data plane adoption].

For a single-cell standalone TripleO deployment:
[,bash]
----
scp -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]:/root/tripleo-standalone-passwords.yaml ~/overcloud-passwords.yaml
----

Further on, this password is going to be referenced as `TRIPLEO_PASSWORDS[default]` for a `default` cell name, in terms of TripleO.

For a source cloud deployment with multiple stacks, change the above command to these:
[,bash]
----
scp -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]:/root/tripleo-standalone-passwords.yaml ~/
scp -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]:overcloud-deploy/overcloud/overcloud-passwords.yaml ~/
scp -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]:overcloud-deploy/cell1/cell1-passwords.yaml ~/
scp -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]:overcloud-deploy/cell2/cell2-passwords.yaml ~/
----
Note that all compute cells of the source cloud always share the same database and messaging passwords.
On the contrary, a generic split-stack topology allows using different passwords files for its stacks.

The development environment is now set up, you can go to the https://openstack-k8s-operators.github.io/data-plane-adoption/[Adoption
documentation]
Expand All @@ -371,8 +482,10 @@ Delete the data-plane and control-plane resources from the CRC vm

[,bash]
----
oc delete --ignore-not-found=true --wait=false openstackdataplanedeployment/openstack
oc delete --ignore-not-found=true --wait=false openstackdataplanedeployment/openstack-nova-compute-ffu
for CELL in $(echo $RENAMED_CELLS); do
oc delete --ignore-not-found=true --wait=false openstackdataplanedeployment/openstack-$CELL
oc delete --ignore-not-found=true --wait=false openstackdataplanedeployment/openstack-nova-compute-ffu-$CELL
done
oc delete --ignore-not-found=true --wait=false openstackcontrolplane/openstack
oc patch openstackcontrolplane openstack --type=merge --patch '
metadata:
Expand All @@ -391,21 +504,33 @@ oc delete --wait=false pod mariadb-copy-data || true
oc delete secret osp-secret || true
----

Revert the standalone vm to the snapshotted state
Revert the standalone vm(s) to the snapshotted state

[,bash]
----
cd ~/install_yamls/devsetup
make standalone_revert
----

For a multi-cell deployment, change the above command to these:
[,bash]
----
cd ~/install_yamls/devsetup
for n in 0 1 2 3 4; do make standalone_revert EDPM_COMPUTE_SUFFIX=$n; done
----

Clean up and initialize the storage PVs in CRC vm

[,bash]
----
cd ..
for i in {1..3}; do make crc_storage_cleanup crc_storage && break || sleep 5; done
for CELL in $(echo $RENAMED_CELLS); do
oc delete pvc mysql-db-openstack-$CELL-galera-0 --ignore-not-found=true
oc delete pvc persistence-rabbitmq-$CELL-server-0 --ignore-not-found=true
done
----
Use indexes like `*-0`, `*-1` based on the replica counts configured in `oscp/openstack` CR.

'''

Expand Down
2 changes: 2 additions & 0 deletions docs_dev/assemblies/tests.adoc
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,9 @@ these variables suit your environment:
** `tripleo_passwords` (for each {OpenStackPreviousInstaller} Heat stack on the source cloud)
** `source_galera_members` (for each cell controller on the source cloud)
** `source_mariadb_ip` (for each cell controller on the source cloud)
** `edpm_nodes` (for each cell compute node on the destination)
** `edpm_privatekey_path`
** `source_ovndb_ip``
** `timesync_ntp_servers`

== Running the tests
Expand Down
Loading

0 comments on commit 6a5c854

Please sign in to comment.