About this Guide
This guide describes how to install a VAST Cluster using the VAST Web UI Cluster Install utility. The Cluster Install utility can help you perform software installation of a VAST Cluster when you are done rack mounting and cabling the hardware components.
Make sure that you are following a proper installation plan under the guidance of your VAST Sales Engineer.
You Will Need
Your Site Survey (completed by a VAST Sales Engineer) and installation plan.
A laptop
The VAST Cluster package file saved to the laptop. For example: release-<x.x.x-xxxxxx>.vast.tar.gz. Contact VAST Support to obtain the package file.
The bootstrap file (
vast_bootstrap.sh) saved to the laptop. Obtain this file from VAST Support.
Run the VAST Web UI Cluster Install Utility
Follow this guide to run the Cluster Install utility after racking and cabling the cluster hardware and configuring the switches. After following the procedure to Connect to a CNode Tech Port, Copy Package, SSH to Management CNode, the installation process proceeds to the Easy Install Utility, which includes the following five screens:
Important
Follow a prepared plan for the specific installation when completing all fields. Specifically, specific input values for your installation should be specified in the VAST Install Wizard Settings tab of your site survey. The field descriptions below are intended as guidance to help you implement a planned configuration.
Connect to a CNode Tech Port, Copy Package, SSH to Management CNode
Configure the Ethernet interface on your laptop to be on the following subnet: 192.168.2.0/24.
Connect your laptop to the technician port on any one of the CNodes. This CNode will become the Management CNode.
Run the following commands to copy the VAST Cluster package file (e.g. release-<x.x.x-xxxxxx>.vast.tar.gz) to the CNode.
scp <package file path> vastdata@192.168.2.2:/vast/bundles/Where
<package file path>is the local path to the package file.Note
Make sure there is only one VAST Cluster package file located at /vast/bundles/.
You'll be prompted for the password on running each command. The default password is vastdata.
Log in to the management CNode via SSH and run the
vast_bootstrap.shscript which is included in VAST OS:username@host:~$ ssh vastdata@192.168.2.2 [vastdata@localhost ~]$ cd /vast/bundles [vastdata@localhost bundles]$ vast_bootstrap.shConfirm the action:
Are you sure you want to reimage? this will wipe the current system [Y/n] Y unpacking release-<x.x.x-xxxxxx>.vast.tar.gz, this may take a whileThe script extracts the package files and runs the VAST Management Server (VMS) container.
When the
vast_bootstrap.shscript is complete, the following message is displayed:bootstrap finished, please connect at https://192.168.2.2While still connected to the technician port, open a web browser on your laptop and browse to https://192.168.2.2.
The VAST Web UI opens and displays the VAST DATA - End User License Agreement.
Click I Agree.
The login page appears.
Log in using the default admin user and password:
Username: admin
Password: 123456
The General Cluster Configuration screen is displayed.
At this stage, the Easy Install utility attempts to discover the CNodes and DNodes that comprise the cluster.
Tip
Nodes are discoverable provided the switches were configured before you began running Easy Install.
General Cluster Configuration
Define the basic information and operations of your cluster on the Cluster Configuration screen.
Select box type to install | Select to install either CBoxes & DBoxes or EBox. |
Cluster name | A name for the cluster. |
PSNT | The cluster's PSNT. A PSNT is an asset identifier that links the components of a cluster. |
License | Enter the license key for the cluster. If no license key is entered, a temporary 30 days license is installed. |
Management network | This field specifies the interface to be used for the management network:
|
B2B IPMI | Depending on the Management Network setting, this can enable the CNodes and/or DNodes for this cluster.
|
DBox HA | Enables NVRAM High Availability (HA) for DBoxes. Support for DBox NVRAM HA is limited. Before enabling this feature, review its usage and limitations. It is possible to enable the feature at a later time after installation, although it will cause a drive layout rewrite. |
B2B template | B2B is a networking configuration option that isolates the IPMI network from the management network. A B2B IP is generated per node as 192.168.3.x, where x is a node index. Optionally, you can set a different B2B template. For example, if you set the B2B template to be 10.250.100 then the B2B IPs will be 10.250.100.x. Default: 192.168.3 |
To enable rack resiliency, toggle Enable Rack Level Resiliency.
To enable NVRAM High Availability (HA) for DBoxes, toggle DBox HA. Support for DBox NVRAM HA is limited. Before enabling this feature, review its usage and limitations. It is possible to enable the feature at a later time after installation, although it will cause a drive layout rewrite.
To enable similarity-based data reduction on the cluster, toggle Similarity. This can also be enabled or disabled after installation.
Note
VAST Cluster encryption of data at rest is FIPS 140-3 capable.
To enable encryption of data at rest on the cluster:
Toggle Encryption.
INTERNAL. Encryption with keys managed internally. This is the only type of encryption that can be disabled after installation or enabled after installation.
CIPHER_TRUST_KMIP. Encryption with keys managed externally on Thales Group CipherTrust Data Security Platform.
FORTANIX_KMIP. Encryption with keys managed externally on Fortanix DSM.
HASHICORP_KMIP. Encryption with keys managed externally on HashiCorp Vault Enterprise.
ENTRUST_KMIP. Encryption with keys managed externally on Entrust KeyControl.
AKEYLESS_KMIP. Encryption with keys managed externally on the Akeyless platform.
If you selected INTERNAL, no further encryption setup is necessary. If you selected an external key manager (EKM) option, complete steps 3-5, as applicable.
Add up to five EKM servers: For each server, enter the server IP address in the Enter the SSL certificate for the connection to the EKM servers.
Enter the SSL certificate for the connection to the EKM servers:
Click the +Add Certificate link under EKM Certificate and then paste the content of the certificate file into the text field provided. Include the "-----BEGIN CERTIFICATE-----" and "-----END CERTIFICATE-----" lines from the certificate file content.
Click the +Add Key link under EKM Private Key and paste the content of the private key file of the SSL certificate into the test field provided.
Optionally, enter a CA certificate: Click the +Add Certificate link under EKM CA certificate and paste the content of the CA certificate file.
For Thales Group CipherTrust Data Security Platform only:
In the Auth Domain field, you can specify a subdomain of the EKM root domain (optional).
The subdomain needs to be created on the Thales CipherTrust manager.
When the deployment is complete, encryption groups created on the cluster will have their encryption-at-rest keys generated within the specified subdomain.
In the Use proxy field, you can specify a proxy server through which to connect to the EKM server (optional). Select the check box and then provide the host and port of the proxy server in the fields provided. Specify Host in the format https://proxy-address.
Enter the first rack in Rack name, selecting +Add Another for each additional rack needed, one at a time.
You can also import a configuration file, using mapped_racks.json. Click Import Configuration File to automatically create mappings between the identified boxes. It also creates an output description of:
physical boxes that were not discovered.
boxes that were discovered but not mapped.
conflicts between manual mapping and the uploaded file mapping.
Issues and errors will be reported based on these points.
Racks can be automatically created when loading the rack box mappings. Box mappings for nodes that were not filled will be automatically created. Nodes that were filled will be skipped.
Click Next to move to Infrastructure Mapping.
Infrastructure Mapping
Assign boxes to rack positions
Expand each box and select every node in the box that you want to include in the cluster.
Note
Boxes or nodes that are discovered but not selected will not be included in the installation.
Rack | To assign the box to a rack, select the rack from the dropdown. Use the identifying details such as the serial number and box type to identify the physical location of the box. |
Unit | Enter a meaningful indicator for the position of the box in the rack. For example, U1 for the lowest unit position in the rack. |
Subsystem | Leave the default value (0) unless the installation requires multiple subsystems. If needed, set the subsystem per DBox per the plan. The subsystem is used in the formation of the IP addresses that are allocated to the nodes for the cluster's internal network. Multiple subsystems expand the number of IP addresses that are available for allocation. The default setting is 0 for all DNodes and CNodes, which configures a single subsystem. A single subsystem enables the allocation of up to 254 IP addresses. There are three internal IP addresses allocated to each CNode and to each DNode. Follow the installation plan for the cluster and allocate a subsystem to each node as planned. Valid range: 0-63. |
Click Next.
Network Configurations
Configure network settings and connectivity for your cluster. You can set these configurations globally or per rack. If you select Global Configuration and Apply selection, the configuration values are filled out only once for the whole cluster. If you select Per Rack Configuration, you must apply the values to each rack in the Rack window.
Global Configurations | |
Racks | |
IPv4 CIDR | The IPv4 mask for the management subnet in CIDR notation. |
IPv6 CIDR | The IPv6 prefix length for the management subnet. |
External gateway | The IPv4 or IPv6 address of the default gateway for the management network. Click within the field or choose Expand to display an IP address entry dialog. Enter the IP address and click +Add. The entry is added to the IPV4 or IPV6 list respectively. Click Save Changes to close the dialog. |
IPMI netmask | The subnet mask for the IPMI default gateway. |
IPMI gateway | The IP of a gateway for the IPMI interfaces on the CNodes and DNodes, if different from the management network default gateway. Example: 173.30.200.1 |
DBoxes | |
DBox Management network | This field specifies the interface to be used for the management network:
|
CBoxes | |
CBox External network type | This field specifies the interface to be used for the management network:
|
NICs Configurations | |
Skip NIC (field shown if applicable) | If CNodes are dual-NIC CNodes and have NICs that are not in use (not connected to any network), use this field to specify which NIC is not connected on each CNode and hence should not be included in the network configuration. Select EXT if you want to use only one of the two NICs. Otherwise, select No, which is the default, for all deployments. |
Reverse NICs | Reverse NICs should be disabled for all deployments. This is also the default. |
Rack | |
Mgmt.IP pool | This section refers to network configurations for each of the racks. The IP pool from which to assign IPs for the management network to all CNodes or EBoxes in the rack. The pool should contain enough IPs for all CNodes or EBoxes in the rack. For example, for an installation with one CBox, there are four CNodes, so you need to supply four IPs that were designated for the management external IP pool in the installation plan. The recommendation "You should add exactly 4 IPs" is displayed. |
IPMI.IP pool | An IP pool from which to assign an IP to the IPMI interface of each CNode or EBox. Set this IP pool if and only if the planned deployment uses the standard IPMI network configuration. If you are deploying B2B IPMI networking, do not configure this IP pool. Instead, configure a B2B template on the General Cluster Configuration page. The CNodes or EBoxes in the rack will be assigned IPMI IPs in the same order as they are assigned management external IPs. The CNode that receives the first IP in the management external IP pool receives the first IP in the CNodes IPMI pool and so on. The node allocation order can be changed by dragging each node manually, or you can select an automatic sort order with the IP Allocation Order selections in the upper right corner of the racks table. You can select ascending or descending sort for either CBoxes or DBoxes. |
After the IP pools are added to the racks, define the order of installation. For each rack, click the arrow to see the nodes. Node order can be changed by dragging the Initial IP column in the list of boxes.
Advanced Settings
Configure advanced cluster settings.
Out of Band Network | |
VMS Management VIP | A virtual IPv4 or IPv6 address configured on the management interfaces on all CNodes. VAST Management System (VMS) listens on this IP. The IP should be on the management subnet. Click within the field to display an IP address entry dialog. Enter the IP address and click +Add. The entry is added to the IPV4 or IPV6 list respectively. Network Configurations CIDR and GW are relevant for this IP as well |
DNS IPs | The IPv4 or IPv6 address(es) of any DNS servers that will forward DNS queries to the cluster. Click within the field or choose Expand to display an IP address entry dialog. Enter the IP address(es) and click +Add. The entry or entries are added to the IPV4 or IPV6 list respectively. Click Save Changes to close the dialog. |
DNS Search Domains | Enter the domains on your data network on which client hosts may reside. If you provide these, you will be able to specify hosts by name instead of IP when setting up export policies, call home settings, webhook definitions and so on. VAST Cluster will use these domains to look up host IPs on the DNS server. |
NTP Server | Changes the server addresses that the cluster will use for Network Time Services. This can be a valid FQDN. |
CNode management external VLAN | Sets a VLAN on the CNode management network external interfaces. |
Customer IP | Checks if there is an external connection to this IP. |
Technician IP | Changes the IP configured on the technician interface. Default: 192.168.2.2 |
Inner Network | |
Data Vlan | For Ethernet configurations, enter the VLAN to isolate the cluster's internal network from the data network. In case of a conflicting use of the default VLAN, enter a different VLAN that is not already used on the client network. Default: 69 Example: 69 |
Management Inner VIPs | Sets the inner VIP of the management. Example: 172.16.254.254 |
Inner IP Template | Sets a custom subnet for the cluster's internal network. Default: 172.16 The Data VLAN isolates the internal network from the external network. If you anticipate IP address collisions with the default subnet, such as in an IB configuration, you can set a custom subnet. Each CNode and DNode is allocated three IP addresses for three networks within the subnet. These are generated within the subnet from a combination of:
The IPs for these interfaces are generated on the nodes as subnet.subsystem.x, where x is an index per node. For example, if the subnet is 10.200, with the default subsystem, data netmask 255.255.192.0 and start indexes, the following IPs are generated for the internal network interfaces on the first DNode: 10.200.0.100, 10.200.64.100, 10.200.128.100. The following IPs are generated for the internal network interfaces on the first CNode: 10.200.0.1, 10.200.64.1, 10.200.128.1. IPs for the equivalent interfaces for subsequent CNodes and DNodes are incremented from these. The subnet mask for the internal network is 255.255.192.0. Each DNode and CNode is configured with three interfaces on the network. |
CNode Start index | Sets the start value for the indexes appended to internal IPs for the CNodes (see also Subnet). Default: 1 |
DNode Start index | Sets the start value for the indexes appended to internal IPs for the DNodes (see also Subnet). Default: 100 |
Docker IP | Specifies a Docker bridge IP (used internally) in case it needs to be changed from the default due to IP conflicts. Default: 172.17.0.1 |
Docker CIDR | Specifies a Docker bridge IP subnet as a CIDR index in case it needs to be changed from the default due to IP conflicts. Default: 16 |
IB MTU (field shown if applicable) | For dual-NIC CNodes where a NIC is directly connected to an external InfiniBand data network, use this field to set the MTU for that InfiniBand network. Default: 2044 Take care to set a supported MTU for the NIC mode:
|
IPoIB mode (field shown if applicable) | Displays the type(s) of external InfiniBand network(s) that the CNode is connected to: Datagram or Connected (default).
|
ETH MTU (field shown if applicable) | For dual-NIC CNodes where a NIC is directly connected to an external Ethernet network, use this field to set the MTU for that Ethernet network. |
Hostname Template | Gives full control over the template of the hostname as needed. Modify the text with placeholders for different parameters that the system will inject. For example:
- ...so after the user puts them - the system will generate the hostname by filling the placeholders with the actual values. Important note - |
Network QoS
Sets the QoS flow control type to run on Mellanox interfaces. The QoS policies can be specified separately for internal and external networks. However, if only one NIC is installed, the internal configuration is used.
Global Pause - Flow control mechanism that prevents packet drops by temporarily halting all traffic across an entire network link when the receiving port's buffers are nearing capacity.
Do not select Global Pause if you are deploying Cumulus switches. Using this option with Cumulus switches can cause significant performance issues.
Note
A system that was configured with Priority Flow Control cannot be changed back to Global Pause.
Priority Flow Control - The network Quality of Service mechanism that prevents data loss during periods of congestion by selectively pausing specific priority classes of traffic rather than halting the entire network link.
Logical Structure
Internal logical configurations of the cluster setup such as number of shards and defrag configurations.
Warning
Do not change settings unless guided to do so by VAST Support.
Installation Flags
Dedicated installation flags that can be provided to the system for special configurations, including vsettings.
Warning
Do not change settings unless guided to do so by VAST Support.
Call Home/Support
Select the Call Home tab and make settings as needed.
The Call Home feature sends non-sensitive data from your VAST Cluster to the VAST support server to enable us to provide proactive analysis and fast response on critical issues. The collected data is sent by HTTPS to a VAST Data AWS S3 bucket that we maintain for this purpose.
Supported Drive Disks
You can upload an updated supported drives file and its signature during the installation in case there is a new drive that was approved/qualified after the VAST Cluster release version date.
For more information, see Drive Compatibility Validation.
Pre-Validation
After completing the configuration screens, you have an option to verify the setup of the cluster. To do so, click Pre-Validation. The process collects all issues and generates a list, aggregated by issue type, which can be viewed and exported to a spreadsheet. Each issue has a suggested action and a severity to allow you the opportunity to fix any issues that would block the installation.
If failures are reported, you can remove components to continue the installation process.
The pre-validation is an optional step. Alternatively, you can go straight to installation. Each installation runs the initial pre-validation process, then continues to the installation if there are no errors.
Installation
Review the settings you made and ensure that they match the installation plan.
If the installation fails at any stage, the cluster_deploy task state shows Failed and the details of the failure are displayed.
The Cluster Install dialog also reopens to enable you to choose whether to resume or restart the installation.
Note
If the dialog is closed, you can open it by clicking the
button at the top of the page to open it.If this occurs:
Check the logged details of the failure and work to resolve the cause of the failure.
When you have resolved the cause of failure, choose one of the following:
To resume the installation from the last successfully completed step before the failure, click Resume Installation.
The installation resumes and you can follow the task in the Activities page again. The Activities page shows you which steps are being skipped as the task progresses.
If any changes to any of the installation parameters are needed, or if you need to replace any component devices, you must start over. To wipe the completed steps and start the installation from the beginning, click Start Over.
When the installation is done, the cluster_deploy task state changes to COMPLETED and the cluster status displayed at the top left of the page changes to Online:
.Check the log details for failed CNodes that may have been skipped during the installation (the installation continues if the limit of skipped CNodes is not reached). These failed CNodes are not included in the cluster. When the issues causing these failures are resolved, you can add these units to the cluster while after it is running, in the VAST Web UI, using Add CNode, in CNode tab of the Infrastructure page.
You can now disconnect from the technician port. The cluster's VAST Web UI is now accessible by browsing to the configured management VIP from network locations that have network access to the management VIP.
To begin managing the cluster, browse to the management VIP and log in using the default username admin and password 123456.