Skip to content

Latest commit

 

History

History
125 lines (88 loc) · 10.1 KB

hana-ha.md

File metadata and controls

125 lines (88 loc) · 10.1 KB

HANA-HA

This stack builds HANA HA instances.

Architecture Diagram

HANA-HA

Prerequisites

Infrastructure

If not using the provided Terraform, the following infrastructure components must exist.

HANA

Two HANA machines are required, one for the primary and one for the secondary.

Disks

Two disks must be created and attached to each HANA machine.

data - This disk will contain four logical volumes, data, shared, log, and usr (for /usr/sap). Attach the disk to the machine with a device_name of data to use the value Ansible uses for it by default, otherwise it requires redefining the variable sap_hana_disks. See the GCP SAP HANA planning guide for disk sizing.

backup - The backup disk will have a single logical volume. Attach the disk to the machine with a device_name of backup to use the value Ansible uses for it by default. See the GCP SAP HANA planning guide for disk sizing.

Load balancer

First, there must be one unmanaged instance group (UMIG) for the primary instance and one for the secondary instance.

An internal load balancer (ILB) is then created with two backends: one whose group is the primary UMIG and another whose group is the secondary UMIG. The primary backend must have failover set to false, and the secondary backend must have failover set to true. The backends should have a failover_policy with disable_connection_drain_on_failover set to true, drop_traffic_if_unhealthy set to true, and failover_ratio set to 1.

Inventory

If using Terraform and Ansible together, the inventory is automatically generated. If using Ansible separately from or without Terraform, the inventory must be defined as shown.

Inventory groups

The following inventory groups must be defined containing the hosts described below. You can choose your own names for the groups.

  • HANA group
    • 1 Primary host
    • 1 Secondary host

HANA inventory host variables

Name Description Type Default Required
sap_hana_is_primary Whether or not this is the primary host. bool n/a yes

Examples

INI format:

# Note: when using INI formatted inventory, boolean inventory values
# must be in uppercase or Ansible will convert them to strings.

[hana]
abchanapri sap_hana_is_primary=True
abchanasec sap_hana_is_primary=False

YAML format:

all:
  children:
    hana:
      hosts:
        abchanapri:
          sap_hana_is_primary: true
        abchanasec:
          sap_hana_is_primary: false

Install Media

See the instructions for uploading install media to your bucket.

Variables

Variables related to Terraform

The following variables are only used when Terraform and Ansible are run together.

Name Description Type Default Required
sap_hana_instance_name Base name of HANA instances. The names of the primary and secondary instances will be automatically derived from this. To set explicit names for the primary and secondary instances, define sap_hana_instance_name_primary and sap_hana_instance_name_secondary instead. string n/a yes, if sap_hana_instance_name_primary and sap_hana_instance_name_secondary are not defined
sap_hana_instance_name_primary Name of primary HANA instance. Use instead of sap_hana_instance_name to explicitly set the name of the primary instance. string n/a yes, if sap_hana_instance_name is not defined
sap_hana_instance_name_secondary Name of secondary HANA instance. Use instead of sap_hana_instance_name to explicitly set the name of the secondary instance. string n/a yes, if sap_hana_instance_name is not defined
sap_hana_instance_type The GCE instance type for HANA. Must be one of n1-highmem-32, n1-highmem-64, n1-highmem-96, n2-highmem-32, n2-highmem-48, n2-highmem-64, n2-highmem-80, m1-megamem-96, m1-ultramem-40, m1-ultramem-80, m1-ultramem-160, m2-ultramem-208, or m2-ultramem-416. string n1-highmem-32 no
sap_hana_service_account_name The name of the service account assigned to HANA instances. This should not be a full service account email, just the name before the @ symbol. string sap-common-sa no
sap_project_id The project ID where instances are located. string n/a yes
sap_primary_zone The zone for primary instances, HANA and ASCS and PAS, for example us-central1-a. string n/a yes
sap_secondary_zone The zone for primary instances, both HANA and ERS, for example us-central1-b. string n/a yes
sap_subnetwork_project_id The name of the subnetwork project, if using a shared VPC. If not given, sap_project_id will be used. string value of sap_project_id no
sap_source_image_family The source image family for machines. string n/a yes
sap_source_image_project The project for the source image. Official SAP images are from rhel-sap-cloud for RedHat or suse-sap-cloud for Suse. string n/a yes
sap_tf_state_bucket The GCS bucket where Terraform state is stored. If it does not exist, it will be created. There can only be one bucket globally with a given name (it gets a global DNS name). If there is a permissions error when creating this bucket, it is likely that one already exists in another project with the same name. Note that the pair sap_tf_state_bucket, sap_tf_state_bucket_prefix must be unique to avoid conflicts with other stacks. string n/a yes
sap_tf_state_bucket_prefix This is the prefix for the Terraform state within the bucket defined in sap_tf_state_bucket. Note that the pair sap_tf_state_bucket, sap_tf_state_bucket_prefix must be unique to avoid conflicts with other stacks. string n/a yes

Additional Variables

The following variables are used with and without Terraform.

Name Description Type Default Required
sap_hana_backint_bucket_name Name of bucket used by backint. string '' no
sap_hana_backint_install Whether or not to install backint. bool n/a yes
sap_hana_fast_restart Whether or not to enable HANA fast restart. bool false no
sap_hana_instance_name_primary Name of primary HANA instance. string n/a yes, if terraform is not used
sap_hana_instance_name_secondary Name of secondary HANA instance. string n/a yes, if terraform is not used
sap_hana_instance_number Instance number for HANA. This is a two digit number that must be in quotes, or Ansible will convert it into single digits, for example 00 without surrounding quotes gets converted to the number 0. string 00 no
sap_hana_install_files_bucket Bucket where HANA installation files are located. string n/a yes
sap_hana_password The password for HANA. string n/a yes
sap_hana_product_version The version of HANA. string 20SPS03 no
sap_hana_sid The System ID for HANA. This is a three character uppercase string which may include digits but must start with a letter. string n/a yes
sap_hana_backup_size The size of the backup volume's filesystem in LVM format. See the GCP SAP HANA planning guide for partition sizing. string n/a yes, if terraform is not used
sap_hana_data_size The size of the data volume's filesystem in LVM format. See the GCP SAP HANA planning guide for partition sizing. string n/a yes, if terraform is not used
sap_hana_log_size The size of the log volume's filesystem in LVM format. See the GCP SAP HANA planning guide for partition sizing. string n/a yes, if terraform is not used
sap_hana_shared_size The size of the shared volume's filesystem in LVM format. See the GCP SAP HANA planning guide for partition sizing. string n/a yes, if terraform is not used
sap_hana_usr_size The size of the /usr/sap volume's filesystem in LVM format. See the GCP SAP HANA planning guide for partition sizing. string n/a yes, if terraform is not used
sap_hana_virtual_host The hostname given to the HANA load balancer, added to /etc/hosts of instances. string n/a yes
sap_hana_vip The IP address of the HANA load balancer, added to /etc/hosts of instances. string n/a yes