IBM Cloud Docs
Modifying a Power Virtual Server instance

Modifying a Power Virtual Server instance

Learn how to modify your Power® Virtual Server to better meet your business needs.

Resizing an instance by using the Power Virtual Server user interface

To resize a Power Virtual Server instance after its initial creation, complete the following steps:

  1. Go to Virtual server instances in the Power Virtual Server user interface and click your instance.

  2. Click the Edit details in the server details pane.

  3. A menu appears after you click the Edit icon. Here, you can change your instance's Name, Entitled capacity, Memory (GB), and VM pinning state.

    If the VM is inactive or shut down, you can change the processor type to Dedicated processor, Uncapped shared processor or Capped shared processor.

  4. Check the service agreement box and click Order to complete the instance modification process and accept the price.

  5. View the Server details pane to verify your instance modification.

Resizing the virtual machine core count and memory

You can scale up and scale down the core count and memory of the virtual machine (VM) as per your workload requirements. When the VM is active, you can resize the memory and core counts to a maximum of eight times of the specified values, and to a minimum of 1/8 times of the specified values when the VM was provisioned. However, you cannot resize the memory and core count to less than 0.25 cores and 2 GB memory. You can resize the memory and core count beyond the 8x and 1/8x boundaries when the VM is shut down. The following table shows an example of recalculated values:

Table 1. Resizing VM core count and memory when the VM is active
Specified value when the VM instance was provisioned Minimum resize value (must be greater than or equal to 0.25 cores, 2 GB memory) Maximum resize value
4 core and 8 GB memory 0.5 cores and 2 GB memory 32 cores and 64 GB memory
Table 1. Resizing VM core count and memory when the VM is shut down
Specified value when the VM instance was provisioned Minimum resize value (must be greater than or equal to 0.25 cores, 2 GB memory) Maximum resize value
4 core and 8 GB memory You can specify any value that is greater than 0.25 cores, 2 GB memory You can specify any value that is smaller than the available resources in the host

If you want to resize an existing VM that was created before 15 December 2020 to 8x ratio of core count and memory, you must first shut down the VM, resize the VM, and then activate the VM. You must resize the VM at least once when the VM is shut down to enable 8x ratio. Simply shutting down and activating the VM does not enable the 8x ratio of core count and memory.

Managing your storage volumes

Learn how to add new storage volumes and modify existing ones. You can now attach storage volumes to a VM instance from different storage tiers and pools, other than the storage pool the VM instance's root (boot) volume is deployed in. To accomplish this you must modify the VM instance and set the new VM instance storagePoolAffinity property to false. You can now attach mixed storage to a VM instance. For more information, see How to set a VM instance to allow attaching mixed storage?.

You can create a storage volume by specifying any name of your choice. If you want to reuse the storage volume name, you must delete the existing storage volume with the same name. After you delete the original volume, allow one hour before creating a new volume with the same name.

Adding and managing storage volumes

If you want to attach or detach a volume, complete the following steps:

  1. Go to Virtual server instances in the Power Virtual Server user interface and click your instance.

  2. Under Attached volumes section, click Attach volume to add a storage volume from the list.

  3. To detach a storage volume, click Detach in the table.

    The user interface may display a failure message when you attempt to detach a volume from a virtual server instance. In such cases, you need to reload the page after a brief delay in order to see if the volume has been successfully detached. Another detach request should be made if the volume is still not disconnected.

  4. You can also create a new storage volume.

  5. Enter the Name, Tier, Number of volumes, and Size of the storage volume. You can also toggle the Shareable switch to the On position to allow multiple virtual instances to write data to the same data volume.

  6. Select one of the following Storage pool options:

    • Auto-select pool: Use this option to allow the system to automatically select a storage pool, for the desired storage tier, that has sufficient capacity.
    • Affinity: Use this option to select an existing VM instance (VM) or an existing volume as the affinity object. The new volume is created in the same storage pool where the affinity object resides. If you are using VM instance as an affinity object, the storage pool that is selected is based on the PMV instance's root (boot) volume.
    • Anti-affinity: Use this option to specify one or more existing VM instances or one or more volumes as the anti-affinity objects. The new volume is created in a different storage pool than the storage pool where one or more anti-affinity objects reside.

    For more information about affinity and anti-affinity policy, see What does it mean to set an affinity or anti-affinity rule?.

    In the API for create volume feature two new properties antiAffinityVMInstances and antiAffinityVolumes have been added. These properties are used only to specify anti-affinity objects. You can specify only one object type for affinity or anti-affinity objects, either VM Instances or Volumes. For more information about storage volumes APIs, see Create a new data volume and Create multiple data volumes from a single definition.

  7. Click Create and Attach.

Resizing a storage volume

You can resize a storage volume after its initial creation. To delete a volume, its status must indicate one of the following states: available, error, error_restoring, error_extending, or error_managing. Additionally, the storage volume cannot be deleted if it is migrating, attached, belongs to a group, has snapshots, or is disassociated from its snapshots after a transfer. Resizing is not immediately available after you deploy a VM. For IBM i 7.3 and later versions, you can resize volume to increase the volume size, but this requires an initial program load (IPL) to recognize the new volume size.

Before you perform the IPL operation, you must run the macro to ensure that the volume resize operation is complete, then proceed with the IPL operation. For more information, see Dynamically increasing the size of a SAN LUN. If you perform an IPL operation before the resize operation is complete, an additional IPL is required.

If you cannot take the downtime, you can add additional volumes. You can attach maximum of 127 volumes to the VM.

Any volume that has been included in a snapshot cannot be resized. To resize a volume that has been included in a snapshot, you must first delete all of the snapshots the volume is a part of.

  1. Go to the Power Virtual Server user interface and click Storage volumes.

  2. Click the Edit icon to the right of your storage volume.

  3. Click on Edit to select the desired storage volume size in the Modify storage volume window. You can increase only the size of the storage volume.

  4. Read the service agreement and agree to the terms. Click Order to complete the volume modification process and accept the price.

  5. To verify your new storage size, go back to Storage volumes.

  6. In an AIX VM instance, if you resize your boot storage volume, run the chvg -g rootvg command.

To apply or verify an IBM i software key, the VM must be active and in running state. If you already ran a resize operation, you must wait until the resize operation complete and VM returns to OK status.

Adding or removing a public network

You can remove or add a public network by clicking the Public networks toggle in Virtual server instances. When you toggle a public network off and then on, the Power Virtual Server user interface regenerates new internal and external IP addresses. You need to check the Power Virtual Server user interface for the new internal IP address (that maps to the external IP address). You must add a network interface controller (NIC) and point it to the new internal IP address. For information on how to add or remove an interface, see How to add or remove a network interface from an AIX virtual machine (VM) or How to add or remove a network interface from an IBM i virtual machine (VM).

You cannot toggle a public network off if there are no other defined networks.

Detecting problems by using the System Reference Code (SRC)

A system reference code (SRC) is a set of eight alphanumeric characters that identifies the name of the system component that detects, the error codes, and the reference code that describes the error condition. When the Power Virtual Server instance detects a problem, an SRC number is displayed along with a timestamp in the Server details page. You can use the SRC to resolve the issue yourself. If you are contacting support to resolve a problem, the SRC number might help the hardware service provider better understand the problem and to provide the solution.

For an IBM i VM, the SRC number can be progress code, operation code, or software code. For more information, see the System Reference Code list in IBM i documentation. For AIX VM instances, the SRC numbers are progress codes that provide information about the stages involved in powering on and performing initial program load (IPL). AIX SRCs refresh once in 2 minutes. For more information, see AIX IPL progress codes.

Use cases

The different scenarios that you may face while requesting for resizing a Power Virtual Server instance's memory are listed below.

You request for resizing both memory and CPU fails

When you attempt to resize the memory as well as CPU of a deployed virtual server instance through a single request it may fail due to the following reasons:

  • There is no free memory available on the host.
  • There is no free memory available on the logical partition as the resources on it are running.
  • The free memory available on the logical partition is less than that of the desired value indicated in the resizing request.
  • You have made multiple attempts for resizing.
  • Currently, there is no preference for memory or CPU on what should be resized first. If the first request being processed fails, the second one will also fail automatically.

Example: When the currently allocated memory for the logical partition is 4GB and you are trying to reduce the value to 2GB and the logical partition at the time of request does not have free 2GB memory for resizing (considering the logical partition is using upto 3.2 GB for running resources in it), then there is a possibility that both the CPU and memory resize will fail.

You request for resizing the memory but you get a partial resize

When you attempt to resize the memory of a deployed virtual server instance through a request, it may partially resize or in the worst scenarios even fail due to the following reasons:

  • There is no free memory available on the host will result in failed request.
  • There is no free memory available on the logical partition as the resources are running on it. This will result in a failed request.
  • The free memory available on the logical partition is less than that of the desired value indicated in the resizing request. This will result in a partial resize.
  • You have made multiple attempts for resizing. This will result in a failed request.

Example:When the currently allocated memory for the logical partition is 4GB and you are trying to reduce the value to 2GB and the logical partition at the time of request does not have free 2GB memory for resizing (considering the logical partition is using upto 3 GB for running resources in it) and can free up only 1 GB, then the partial resize should be possible to reduce the memory to 3GB.

In the current cloud environment, it may take upto 1.5 hours approximately for the change in memory to be updated to places referring to the memory of the logical partition. Hence if you attempt to repeat the resize request, the consecutive retires will fail, until all referencing tables are updated.