Getting started with Trilio on Red-Hat OpenStack Platform (RHOSP)
The Red Hat OpenStack Platform Director is the supported and recommended method to deploy and maintain any RHOSP installation.
Trilio is integrating natively into the RHOSP Director. Manual deployment methods are not supported for RHOSP.
1. Prepare for deployment
Refer to the below-mentioned acceptable values for the placeholders triliovault_tag
, trilio_branch
, RHOSP_version
and CONTAINER-TAG-VERSION
in this document as per the Openstack environment:
Trilio Release | triliovault_tag | trilio_branch | RHOSP_version | CONTAINER-TAG-VERSION |
---|---|---|---|---|
6.0.0 | 6.0.0-beta-2-rhosp17.1 | 6.0.0-beta-2 | RHOSP17.1 | 6.0.0-beta-2 |
1.1] Select 'backup target' type
Backup target storage is used to store backup images taken by Trilio and details needed for configuration:
The following backup target types are supported by Trilio
a) NFS
Need NFS share path
b) Amazon S3
- S3 Access Key - Secret Key - Region - Bucket name
c) Other S3 compatible storage (Like Ceph based S3)
- S3 Access Key - Secret Key - Region - Endpoint URL (Valid for S3 other than Amazon S3) - Bucket name
1.2] Clone triliovault-cfg-scripts repository
The following steps are to be done on the undercloud
node on an already installed RHOSP environment.
The overcloud-deploy command has to be run successfully already and the overcloud should be available.
All commands need to be run as a stack
user on the undercloud
node
Refer to the word <RHOSP_RELEASE_DIRECTORY> as rhosp17 in the below sections
The following command clones the triliovault-cfg-scripts github repository.
1.3] Set executable permissions for all shell scripts
1.4] If the backup target type is 'Ceph based S3' with SSL:
If your backup target is CEPH S3 with SSL and SSL certificates are self-signed or authorized by a private CA, then the user needs to provide a CA chain certificate to validate the SSL requests. For all s3 backup target with self signed TLS certificates, user need to copy ca chain files in following location and in given file name format in trilio puppet module. Edit <S3_BACKUP_TARGET_NAME>, <S3_SELF_SIGNED_CERT_CA_CHAIN_FILE> parameters in following command.
For example if S3_BACKUP_TARGET_NAME = BT2_S3 and your S3_SELF_SIGNED_CERT_CA_CHAIN_FILE='s3-ca.pem' then command to copy this ca chain file to trilio puppet module would be
1.5] If the VM migration feature needs to be enabled:
More about this feature can be found here.
Please refer to this page to collect the necessary artifacts before continuing further.
1.5.1] Copy the VMware vSphere Virtual Disk Development Kit 7.0
Rename the file as vddk.tar.gz
and place it at
1.5.2] For secure connection with the vCenter
Copy the vCenter SSL cert file to
2] Update the overcloud roles data file to include Trilio services
Trilio contains multiple services. Add these services to your roles_data.yaml
.
In the case of uncustomized roles_data.yaml can the default file be found on the undercloud
at:
/usr/share/openstack-tripleo-heat-templates/roles_data.yaml
Add the following services to the roles_data.yaml
All commands need to be run as a 'stack' user
2.1] Add Trilio Datamover Api and Trilio Workload Manager services to role data file
This service needs to share the same role as the keystone
and database
service.
In the case of the pre-defined roles, these services will run on the role Controller
.
In the case of custom-defined roles, it is necessary to use the same role where OS::TripleO::Services::Keystone
service installed.
Add the following line to the identified role:
2.2] Add Trilio Datamover Service to role data file
This service needs to share the same role as the nova-compute
service.
In the case of the pre-defined roles will the nova-compute
service run on the role Compute
.
In the case of custom-defined roles, it is necessary to use the role that the nova-compute
service uses.
Add the following line to the identified role:
3] Prepare Trilio container images
All commands need to be run as a 'stack' user
Trilio containers are pushed to the RedHat Container Registry
.
Registry URL: 'registry.connect.redhat.com'. Container pull URLs are given below.
There are three registry methods available in the RedHat OpenStack Platform.
Remote Registry
Local Registry
Satellite Server
3.1] Remote Registry
Follow this section when 'Remote Registry' is used.
In this method, container images get downloaded directly on overcloud nodes during overcloud deploy/update command execution. Users can set the remote registry to a redhat registry or any other private registry that they want to use.
The user needs to provide credentials for the registry in containers-prepare-parameter.yaml
file.
Make sure other OpenStack service images are also using the same method to pull container images. If it's not the case you can not use this method.
Populate
containers-prepare-parameter.yaml
with content like the following. Important parameters arepush_destination: false
, ContainerImageRegistryLogin: true and registry credentials. Trilio container images are published to registryregistry.connect.redhat.com
. Credentials of registry 'registry.redhat.io' will work forregistry.connect.redhat.com
registry too.
Note: This file containers-prepare-parameter.yaml
Redhat document for remote registry method: https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.1/html/director_installation_and_usage/preparing-for-director-installation#container-image-preparation-parameters
Note: File 'containers-prepare-parameter.yaml' gets created as output of command 'openstack tripleo container image prepare'. Refer above document by RedHat
3. Make sure you have network connectivity to the above registries from all overcloud nodes. Otherwise image pull operation will fail.
4. The user needs to manually populate the trilio_env.yaml
file with Trilio container image URLs as given below:
trilio_env.yaml file path:
cd triliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/
At this step, you have configured Trilio image URLs in the necessary environment file.
3.2] Local Registry
Follow this section when 'local registry' is used on the undercloud.
In this case, it is necessary to push the Trilio containers to the undercloud registry.
Trilio provides shell scripts that will pull the containers from registry.connect.redhat.com
and push them to the undercloud and update the trilio_env.yaml
.
At this step, you have downloaded Trilio container images and configured Trilio image URLs in the necessary environment file.
3.3] Red Hat Satellite Server
Follow this section when a Satellite Server is used for the container registry.
Pull the Trilio containers on the Red Hat Satellite using the given Red Hat registry URLs.
Populate the trilio_env.yaml
with container URLs.
At this step, you have downloaded Trilio container images into the RedHat satellite server and configured Trilio image URLs in the necessary environment file.
4] Provide environment details in trilio-env.yaml
Edit /home/stack/triliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_env.yaml
file and provide backup target details and other necessary details in the provided environment file. This environment file will be used in the overcloud deployment to configure Trilio components. Container image names have already been populated in the preparation of the container images. Still, it is recommended to verify the container URLs.
You don't need to provide anything for resource_registry
, keep it as it is.
Parameter | Description |
---|---|
CloudAdminUserName | Default value is Provide the cloudadmin user name of your overcloud |
CloudAdminProjectName | Default value is Provide the cloudadmin project name of your overcloud |
CloudAdminDomainName | Default value is Provide the cloudadmin project name of your overcloud |
CloudAdminPassword | Provide the cloudadmin user's password of your overcloud |
ContainerTriliovaultDatamoverImage | Trilio Datamover Container image name have already been populated in the preparation of the container images. Still it is recommended to verify the container URL. |
ContainerTriliovaultDatamoverApiImage | Trilio DatamoverApi Container image name have already been populated in the preparation of the container images. Still it is recommended to verify the container URL. |
ContainerTriliovaultWlmImage | Trilio WLM Container image name have already been populated in the preparation of the container images. Still it is recommended to verify the container URL. |
ContainerHorizonImage | Horizon Container image name have already been populated in the preparation of the container images. Still it is recommended to verify the container URL. |
TrilioBackupTargets | List of Backup Targets for TrilioVault. These backup targets will be used to store backups taken by TrilioVault. Backup target examples and format of NFS and S3 types are already provided in the trilio_env.yaml file. Details of respective prameters under TrilioBackupTargets given in next section |
TrilioDatamoverOptVolumes | User can specify list of extra volumes that they want to mount on 'triliovault_datamover' container. Refer the |
4.1] TrilioBackupTargets Details
T4O supports setting up multiple target backend for storing snapshots. User can define any number of storage backends as required. At a high level, NFS and S3 are supported.
Following table provides the details of parameters to be set in trilio_env.yaml
file against all the S3 target backends.
Parameters | Description |
---|---|
backup_target_name | User Defined Name of the target backend. Can be any name which can help in quick identifying respective target |
backup_target_type | s3 |
is_default | Can be |
s3_type | Could be either of |
s3_access_key | S3 Access Key |
s3_secret_key | S3 Secret Key |
s3_region_name | S3 Region name |
s3_bucket | S3 Bucket |
s3_endpoint_url | S3 endpoint url |
s3_signature_version | Provide S3 signature version |
s3_auth_version | Provide S3 auth version |
s3_ssl_enabled | true |
s3_ssl_verify | true |
s3_self_signed_cert | true |
s3_bucket_object_lock_enabled | If S3 bucket is having object lock enabled, then this should be set as |
Following table provides the details of parameters to be set in trilio_env.yaml
file against all the NFS target backends.
Parameters | Description |
---|---|
backup_target_name | User Defined Name of the target backend. Can be any name which can help in quick identifying respective target |
backup_target_type | nfs |
is_default | Can be |
nfs_options |
|
is_multi_ip_nfs |
|
nfs_shares | NFS IP and share path. To be kept in case of single IP NFS. Eg. |
multi_ip_nfs_map | NFS IPs and share paths. To be kept in case of multiple NFS IPs. Sample below multi_ip_nfs_map: controller1: 192.168.2.3:/var/nfsshare controller2: 192.168.2.4:/var/nfsshare compute0: 192.168.3.2:/var/nfsshare compute1: 192.168.3.4:/var/nfsshare |
4.2] Update triliovault object store yaml
After you fill in details of backup targets in trilio_env.yaml, user needs to run following script from ‘scripts' directory on undercloud node. This script will update ‘services/triliovault-object-store.yaml' file. User don’t need to verify that.
4.3] For enabling the VM migration feature, populate the below mentioned parameters
More about this feature can be found here.
Parameters | Description |
---|---|
VmwareToOpenstackMigrationEnabled | Set it to |
VcenterUrl | vCenter access URL
example:
|
VcenterUsername | Access username (Check out the privilege requirement here) |
VcenterPassword | Access user's Password |
VcenterNoSsl | If the connection is to be established securely, set it to |
VcenterCACertFileName | If |
5] Generate random passwords for TrilioVault OpenStack resources
The user needs to generate random passwords for Trilio resources using the following script.
This script will generate random passwords for all Trilio resources that are going to get created in OpenStack cloud.
5.1] Change the directory and run the script
5.2] Output will be written to the below file.
Include this file in your overcloud deploy command as an environment file with the option "-e"
6] Fetch ids of required OpenStack resources
6.1] User needs to source 'overcloudrc' file of cloud admin user. This is needed to run OpenStack CLI.
For only this section user needs to source the cloudrc file of overcloud node
6.2] User must have filled in the cloud admin user details of overcloud in 'trilio_env.yaml' file in the 'Provide environment details in trilio-env.yaml' section. If not please do so.
6.3] Cloud admin user should have admin role on cloud admin domain
6.4] After this, user can run the following script.
The output will be written to
7] Load necessary Linux drivers on all Controller and Compute nodes
For TrilioVault functionality to work, we need the following Linux kernel modules to be loaded on all controllers and compute nodes(Where Trilio WLM and Datamover services are going to be installed).
7.1] Install nbd module
7.2] Install fuse module
8] Upload Trilio puppet module
All commands need to be run as a 'stack' user on undercloud node
8.1] Source the stackrc
8.2] The following commands upload the Trilio puppet module to the overcloud registry. The actual upload happens upon the next deployment.
9] Deploy overcloud with Trilio environment
9.1] Include the environment file defaults.yaml
in overcloud deploy command with `-e` option as shown below.
defaults.yaml
in overcloud deploy command with `-e` option as shown below.This YAML file holds the default values, like default Trustee Role is creator
and Keystone endpoint interface is Internal
. There are some other parameters as well those User can update as per their requirements.
9.2] Additionally include the following heat environment files and roles data file we mentioned in the above sections in an overcloud deploy command:
trilio_env.yaml
roles_data.yaml
passwords.yaml
defaults.yaml
Use the correct Trilio endpoint map file as per the available Keystone endpoint configuration. You have to remove your OpenStack's endpoint map file from overcloud deploy command and instead of that use Trilio endpoint map file.
Instead of
tls-endpoints-public-dns.yaml
file, usetriliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_env_tls_endpoints_public_dns.yaml
Instead of
tls-endpoints-public-ip.yaml
file, usetriliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_env_tls_endpoints_public_ip.yaml
Instead of
tls-everywhere-endpoints-dns.yaml
file, usetriliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_env_tls_everywhere_dns.yaml
Instead of
no-tls-endpoints-public-ip.yaml
file, usetriliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_env_non_tls_endpoints_ip.yaml
To include new environment files use -e
option and for roles data files use -r
option.\
Below is an example of an overcloud deploy command with Trilio environment:
Post-deployment for the multipath enabled environment, log into respective Datamover container and add uxsock_timeout with value as 60000 (i.e. 60 sec) in /etc/multipath.conf. Restart datamover container
10] Verify deployment
Please follow this documentation to verify the deployment.
11] Troubleshooting for overcloud deployment failures
Trilio components will be deployed using puppet scripts.
In case of the overcloud deployment failing do the following command provide the list of errors. The following document also provides valuable insights: https://docs.openstack.org/tripleo-docs/latest/install/troubleshooting/troubleshooting-overcloud.html
=> If any Trilio containers do not start well or are in a restarting state on the Controller/Compute node, use the following logs to debug.
12] Advanced Settings/Configuration
12.1] Configure Multi-IP NFS
This section is only required when the Multi-IP feature for NFS is required.
This feature allows us to set the IP to access the NFS Volume per datamover instead of globally.
i] On Undercloud node, change the directory
ii] Edit file triliovault_nfs_map_input.yml
in the current directory and provide compute host and NFS share/IP map.
triliovault_nfs_map_input.yml
in the current directory and provide compute host and NFS share/IP map.Get the overcloud Controller and Compute hostnames from the following command. Check Name
column. Use exact host names in the triliovault_nfs_map_input.yml
file.
Run this command on undercloud by sourcing stackrc
.
Edit the input map file triliovault_nfs_map_input.yml
and fill in all the details. Refer to this page for details about the structure.
Below is an example of how you can set the multi-IP NFS details:
You can not configure the different IPs for the Controllers/WLM nodes, you need to use the same share on all the controller nodes. You can configure the different IPs for Compute/Datamover nodes
iii] Update pyyaml on the undercloud node only
If pip isn't available please install pip on the undercloud.
Expand the map file to create a one-to-one mapping of the compute nodes and the NFS shares.
iv] Validate output map file
The result will be stored in the triliovault_nfs_map_output.yml
file
Open file triliovault_nfs_map_output.yml
available in the current directory and validate that all compute nodes are covered with all the necessary NFS shares.
v] Append this output map file to triliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_nfs_map.yaml
triliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_nfs_map.yaml
Validate the changes in the file triliovault-cfg-scripts/redhat-director-scripts/<RHOSP_RELEASE_DIRECTORY>/environments/trilio_nfs_map.yaml
vi] Include the environment file (trilio_nfs_map.yaml
) in overcloud deploy command with '-e' option as shown below.
trilio_nfs_map.yaml
) in overcloud deploy command with '-e' option as shown below.vii] Ensure that MultiIPNfsEnabled
is set to true in the trilio_env.yaml
file and that NFS is used as a backup target.
MultiIPNfsEnabled
is set to true in the trilio_env.yaml
file and that NFS is used as a backup target.viii] After this you need to run the overcloud deployment.
12.2] Haproxy customized configuration for Trilio dmapi service
The existing default HAproxy configuration works fine with most of the environments. Only when timeout issues with the Trilio Datamover Api are observed or other reasons are known, change the configuration as described here.
Following is the HAproxy conf file location on HAproxy nodes of the overcloud. Trilio Datamover API service HAproxy configuration gets added to this file.
Trilio Datamover HAproxy default configuration from the above file looks as follows:
The user can change the following configuration parameter values.
To change these default values, you need to do the following steps. i) On the undercloud node, open the following file for editing.
ii) Search the following entries and edit as required
iii) Save the changes and do the overcloud deployment again to reflect these changes for overcloud nodes.
12.3] Configure Custom Volume/Directory Mounts for the Trilio Datamover Service
i) If the user wants to add one or more extra volume/directory mounts to the Trilio Datamover Service container, they can use a variable named 'TrilioDatamoverOptVolumes' is available in the below file.
To add one more extra volume/directoy mount to the Trilio Datamover Service container it is necessary that volumes/directories should already be mounted on the Compute host
ii) The variable 'TrilioDatamoverOptVolumes' accepts a list of volume/bind mounts. User needs to edit the file and add their volume mounts in the below format.
iii) Lastly you need to do overcloud deploy/update.
After successful deployment, you will see that volume/directory mount will be mounted inside the Trilio Datamover Service container.
12.4] Advanced Ceph Configration (Optional)
We are using cinder's ceph user for interacting with Ceph cinder storage. This user name is defined using parameter - 'ceph_cinder_user'.
Details about multiple ceph configuration can be found here.
Last updated