There are several IBM privacy settings that would be of interest to an administrator. We're going to try to cover some of the most important settings and where to find them in this lesson, part one and also in part two. There is a menu option called configuration, which is accessible from the top navigation bar. Most configuration settings can be found here. This is the configuration page as of released version 1.4.4 and may change in future releases. There are expandable sections in this view. The main expandable tabs are user management, virtual machines, storage, and e-mail notifications. If database has a service is enabled in your IBM Cloud PowerVC management system, a separate database section is available as shown here. Placement policies are important settings an administrator would like to control. You can use it to specify how virtual machines are distributed within a host group during deployment. IBM privacy requires that every host B and A host group, there's a default group that cannot be deleted, which is where hosts end up by default. However, you can define host groups to logically group hosts. You can then deploy the EMS to a specific host or to a host group. When deploying to a host group, you can choose from one of six placement policies. They are packing, striping, CPU allocation balanced, CPU utilization balanced, memory allocation balanced, and finally, memory utilization balanced. Let's explore each of these options. First, we have this striping policy, what does it mean? This policy means that IBM PowerVC will distribute your virtual machines evenly across all of your hosts. For each virtual machine deployment, a PowerVC detects which hosts have enough processing units and memory to meet the requirements of the virtual machine. Other factors for determining eligible hosts include the storage and network connectivity that are acquired by the virtual machine. From the group of eligible hosts, PowerVC chooses the host that contains the fewest number of virtual machines and places the virtual machine on that host. This way, as newer virtual machines come into play, they get placed in a different host than the one they used previously. Then we have the packing policy. As you may have guessed, this policy will place the virtual machine on the host with the highest number of virtual machines until it's full, that is in terms of virtual processes and memory, it will then move on to the next fullest host. This policy is useful to densely pack the virtual machines into their hosts. Next, we have the CPU utilization balanced placement policy. This one is easy to understand. It places virtual machines on the host that has the lowest CPU utilization in the host group. The CPU utilization is calculated using a running average over the last 15 minutes. A modification on this is the CPU allocation balanced placement policy. This places virtual machines on the host that has the lowest percentage of its CPU allocated post deploy or relocation. Let's consider an example. Let's say you have two hosts and you want to deploy one new virtual machine, which uses only one processor. Let's say that before deployment, host one, has 16 processes, four of which are currently being used by other virtual machines. The second host, host two, has four total processes, two of which are already assigned to virtual machines. The question is where the new virtual machine will go after deployment, and that's easy to calculate. After deployment, host one would have five out of 16 of processes allocated, and host two would have three out of its four processes allocated. Host one therefore, will have the lowest percentage of CPUs allocated, so the virtual machine will go to host one. We've seen how to place a virtual machine based on the number of virtual CPU cores used. What if you want to place virtual machines based on memory utilization and allocation? Well, that's also possible with PowerVC. Again the memory allocation balanced placement policy places virtual machines on the host that has the lowest percentage of its memory allocated, post deploy or relocation. Let's consider another example. Again, we've got two hosts and you want to deploy one new virtual machine, which this time uses four gigabytes of memory. Let's say that before deployment, host one has 32 gigabytes of RAM, four of which are being used by other virtual machines. The second host, host two, has 64 gigabytes of RAM, four of which are already assigned to virtual machines. The question again, is where the new virtual machine will go after deployment. After deployment, host one would have eight out of 32 gigabytes memory allocated, and host two would have eight out of 64 gigabytes of memory allocated. Host two, therefore will have the lowest percentage of memory allocated, so the virtual machine will go to host two. The next setting we're going to look at is called collocation policy. Collocation rules are used to keep virtual machines together or apart. Affinity means that these machines should be kept together, and anti-affinity means that these machines should never be kept together. The rules are enforced when a virtual machine is relocated. Why do we need this? There are several use cases, specifically in high availability scenarios. For example, you want to force a pair of virtual machines on different physical hosts to avoid a single point of failure. You just tack the machines with an anti-affinity policy, so they are not accidentally migrated to the same host. In this diagram, the green machines have the affinity policy on them, so they are place together. The red ones have anti-affinity policies on them, so they are separated. The blue ones don't have anything set yet, so they won't have any constraints regarding their placement. How exactly do we configure collocation rules? This setting is available under the virtual machines expandable tab under configuration. You can create a collocation rule at anytime and then add, or remove virtual machines at anytime after the virtual machines have been deployed. First we click on the create button, give it a name, and pick the affinity or anti-affinity policy. Then we select the virtual machines that need to be part of this rule. Each virtual machine can belong to one or more collocation rules. During migration and remote restart, IBM PowerVC ensures that these rules are followed. Let's pause here and we'll continue our discussion of PowerVC settings in the next video, starting with templates.
