FAQs for IBM Cloud Load Balancer

How many load-balancing options are available in IBM Cloud?

For a detailed comparison of IBM's load balancer offerings, refer to Exploring IBM Cloud load balancers.

Can I use a different DNS name for my load balancer?

While you can't customize the auto-assigned DNS name for the load balancer, you can add a Canonical Name (CNAME) record that points your preferred DNS name to the auto-assigned load balancer DNS name.

For example, if your account number is 123456, your load balancer is deployed in dal09 data center and its name is myapp, the auto-assigned load balancer DNS name is myapp-123456-dal09.lb.bluemix.net. Your preferred DNS name is www.myapp.com. You can add a CNAME record (through the DNS provider that you use to manage myapp.com) pointing www.myapp.com to the load balancer DNS name myapp-12345-dal09.lb.bluemix.net.

What's the maximum number of virtual ports that I can define with my load balancer service?

While trying to create a load balancer service, you can define up to two virtual ports. You can define extra virtual ports after the service is created. The maximum number of virtual ports that are allowed is 10.

What's the maximum number of compute instances that I can associate with my load balancer?

When creating a load balancer service, you can configure up to 10 compute instances as back-end servers. You can define extra servers after the load balancer is created. The maximum number of back-end members that are allowed is 50.

Can my back-end compute instances sit on a subnet different from the load balancer's subnet?

Yes, the load balancer and the compute instances that are connected to the load balancer can be in different subnets, but VLAN spanning must be enabled for the load balancer to communicate and forward requests to the compute instance. For more information, see VLAN spanning.

What are the default settings and allowed values for various health check options?

The default settings and allowed values are as follows.

• Health check interval - Default is 5 seconds, range is 2 – 60 seconds
• Health check response timeout - Default is 2 seconds, range is 1 – 59 seconds
• Max retry attempts - Default is two retry attempts, and the range is 1 - 10 retries

The health check response timeout value must always be less than the health check interval value.

Can I use compute instances residing in remote data centers with this service?

It is recommended that your load balancer service and your compute instances reside locally within the same data center. The load balancer service’s UI does not show compute instances from other remote data centers. However, the UI includes compute instances from other data centers within the same city (for example, data centers whose names share the first three letters, such as DALxx). You can use the API interface to add compute instances from any remote data center.

Which TLS version does SSL offload support?

The IBM Cloud Load Balancer service supports TLS 1.2 with SSL termination.

The following list details the supported ciphers (listed in order of precedence):

• ECDHE-RSA-AES256-GCM-SHA384
• ECDHE-RSA-AES256-SHA384
• AES256-GCM-SHA384
• AES256-SHA256
• ECDHE-RSA-AES128-GCM-SHA256
• ECDHE-RSA-AES128-SHA256
• AES128-GCM-SHA256
• AES128-SHA256

What is the maximum number of load balancer service instances that I can create within my account?

Currently, you can create up to 50 service instances. If you need more instances, contact IBM Support.

Can the IBM Cloud® Load Balancer service be used with VMWare?

VMWare virtual machines that are assigned IBM Cloud portable private addresses can be specified as back-end servers to the load balancer. This feature is available using the API only, and not the web UI. Portable private IPs added by using the API appear as "Unknown" in the UI as they are not assigned by IBM Cloud. This configuration can be used with other hypervisors, such as Xen and KVM.

VMWare virtual machines assigned non-IBM Cloud addresses (such as VMWare NSX networks) cannot be added directly as back-end servers to the load balancer. However, depending on your configuration, it might be possible to configure an intermediary, such as an NSX gateway, that has a IBM Cloud private address as the back-end server to the load balancer (with the actual servers being VMs attached to networks managed by VMware NSX).

If I choose to use a public VLAN under my account to deploy my load balancer and I have a firewall deployed on my public VLAN, what configurations are required on my firewall to work with my load balancer service?

TCP port 56501 is used for management. Ensure that incoming traffic to this port is not blocked by your firewall. Otherwise, load balancer provisioning, as well as customer and service triggered operations, may fail. More specifically, ports 56501 (management), 443 (monitoring), 8834 and 10514 (security and compliance) must be allowed at all times for the load balancer to successfully manage customer workloads. Some outbound traffic is also required to be open to make sure the load balancer functions properly.

In summary, this is the required firewall configuration:

Also, ensure your application's ports are open to accept traffic.

What if I cannot see the monitoring metrics of an existing load balancer after linking my Softlayer account to IBM Cloud?

Monitoring metrics are not available for existing load balancers after linking the accounts. Re-create the load balancers or contact IBM Support. Monitoring metrics for newly created load balancers is available.

Are the load balancer IP addresses fixed?

IBM cannot guarantee load balancer IP addresses to remain constant due to the elasticity that is built into the service. As it scales up or down, you see changes in the available IPs associated with the FQDN of your instance.

Use FQDN and not cached IP addresses.

If I have a firewall deployed in front of my backend servers, what configurations are required for it to work with my load balancer?

The available range of possible IPs for public to public load balancers cannot be predicted. Because of this, you should open all back-end member ports that have been added to the load balancer and set the source IP to any.

Public to private and private to private type load balancers communicate with your back-end members from your own private subnets. Because of this, you can set the source IP with your subnet's CIDR. Note that if the data center where you created the load balancer is part of an MZR, one load balancer appliance deploys in the selected data center, while a second deploys in a different data center within the same region. This means that they exist in two different subnets.

Can the IBM Cloud Load Balancer service be used with Terraform?

Terraform can be used to create, update, and delete an IBM Cloud Load Balancer service resource.

Is it possible to add members with secondary IP addresses?

Members with secondary IP addresses can be added by using the API.

What happens when the load balancer is in Maintenance Pending state?

The load balancer can go into Maintenance Pending state due to following reasons:

• A horizontal scaling event takes place.
• The load balancer has a firewall that's blocking the management traffic.
• Security patches are being installed.

When the load balancer is in the Maintenance Pending state, the data path is not affected.

What is a nonsystem pool?

A nonsystem pool is applicable only with public to private load balancers. The public IP addresses of the load balancer appliances are allocated from your public subnet. Select the Allocate from a public subnet in this account option when provisioning a load balancer.

Does IBM Cloud Load Balancer support UDP?

The IBM Cloud Load Balancer service does not support UDP. It supports only TCP, HTTP, and HTTPS.

Is autoscaling supported by IBM Cloud Load Balancer?

The IBM Cloud Load Balancer service does not support autoscaling currently.

How can I monitor my IBM Cloud Load Balancer metrics?

The IBM Cloud Load Balancer service supports monitoring with IBM Cloud Monitoring. For more information, see Monitoring metrics that use IBM Cloud Monitoring.

What information do I need to file an IBM Support ticket?

To file an IBM Support ticket, provide the product name ("IBM Cloud® Load Balancer"), the UUID of your load balancer (if possible) and your IBM Cloud account number. The UUID can be found in the URL after navigating to the overview page of the given load balancer.

Where can I find pricing information for IBM Cloud Load Balancer?

Pricing metrics for IBM Cloud Load Balancer are detailed in this topic. You can estimate the cost of a service by using the cost estimator on the provisioning pages for IBM Cloud Load Balancer. Select IBM Cloud® Load Balancer from the Load Balancer page of the IBM Cloud catalog, then click Create.

Can I use my Classic Bandwidth Pool to cover bandwidth costs for my IBM Cloud Load Balancer?

No. Currently, IBM Cloud Load Balancer is not eligible to participate in Classic Bandwidth Pools.

Do I need extra IPs in the subnet for IBM Cloud load balancer operations?

It is recommended that you allocate 8 extra IPs per subnet to accommodate horizontal scaling and maintenance operations. If you provision your application load balancer with one subnet, allocate 16 extra IPs.