02-03-14

VCAP-DTA – Objective 3.1 – Configure Pool Storage for Optimal Performance

So this objective sees us moving into section 3 which is entitled “Deploy, Manage, and Customize Pool Implementations”. This objective deals with how we use storage tiers for different virtual disks and use cases, and the sub settings within them. So as usual, let’s run through the skills and abilities for this objective :-

  • Implement storage tiers – When creating a Composer based pool, select the option in the Storage Optimization wizard screen to separate out disks to different datastores. Depending on the exam scenario, you may be asked to separate the Persistent Disks and/or the Replica Disks. Depending on what you select, when you click Next you will get a differing set of options. Assuming you select both, on the vCenter Settings screen, use options 6, 7 and optionally 8  to choose which datastores are used and for which purpose. Once you have completed your choices, complete the wizard out to create the pool.
  • Optimize placement of replica virtual machine – The replica disk is the disk that gets hammered for read read requests from users, so you will be asked to place this on high performance storage, most likely SSD. Using the steps detailed above, use the vCenter Settings screen of the pool wizard to choose a high performance datastore for the replica disk. The diagram below illustrates this point.

replica-ds

  • Configure disposable files and persistent disks – Again this is selected in the pool wizard. You can see from above that there is a View Composer Disks section. This defines how disposable (so think temp files) and persistent disk (user profile) are handled. So for the Persistent Disk, you can select a disk size and drive letter and to redirect the user profile to this disk. The same goes for the Disposable Disk, select the size, whether or not to redirect and which drive letter to use. See below for an illustration of this.

composer-disks

  • Configure and optimize storage for floating or dedicated pools – This is pretty much covered by the first section, Implement Storage Tiers.
  • Configure overcommit settings –  This setting is used when using View Composer. The purpose of overcommit is to allow more disks to be created than physical space exists on the datastore. This is because the disks are sparse disks  on the datastore. The choices for overcommit are None (x0), Conservative (x4, default), Moderate (x7) and Aggressive (x15).  Select the datastore and choose the level of overcommitment from the drop down menu. These choices are only available for OS and Persistent Disks. See below for an example of the dialog.

overcommit

  • Determine implications of using local or shared storage – So in most cases you will be looking to use shared storage, but there may be occasions (and exam scenarios) where you will be asked to use local storage (or it’s use is implied by the question). Bear the following in mind from the View Administration Guide :-
    • You cannot load-balance virtual machines across a resource pool. For example, you cannot use the View Composer rebalance operation with linked-clones that are stored on datastores
    • You cannot use VMware High Availability
    • You cannot use the vSphere Distributed Resource Scheduler (DRS)
    • You cannot store a View Composer replica and linked clones on separate datastores if the replica is on a local datastore
    • When you store linked clones on datastores, VMware strongly recommends that you store the replica on the same volume as the linked clones. Although it is possible to store linked clones on local datastores and the replica on a shared datastore if all ESXi hosts in the cluster can access the replica, VMware does not recommend this configuration
    • If you use floating assignments and perform regular refresh and delete operations, you can successfully deploy linked clones to local datastores.
  • Configure View Storage Accelerator and regeneration cycle – The View Storage Accelerator is also known as the Content Based Read Cache (CBRC) on the ESXi host. This is especially useful as common read based requests are cached into host RAM and is useful for use cases such as desktop boot storms. Configuration is pretty simple – in the pool creation wizard you make your choices in the Advanced Storage Options screen. Check the box to Use View Storage Accelerator, choose between OS Disks  or OS and Persistent Disks. The default is OS disks as this is the usual use case. You also have the option to set a default value for Regenerate Storage Accelerator after days. This basically creates new indexes of the disks and stores them in the digest file for each VM. It’s also worth noting you can configure blackout periods when storage accelerator regeneration will not be run. An obvious example is to suspend this during backups. You may be asked this in the exam. See below for an example.

cbrc

Advertisements

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s