Press "Enter" to skip to content

Category: Virtualization

How VMware Resource Pools Affect SQL Server

Published 2020-04-07 by Kevin Feasel

David Klee walks us through the concept of resource pools in VMware:

Resource pools are used to hierarchically partition available CPU and memory resources, and are available for use at the VMware host cluster layer.

To better prioritize certain VMs over others, especially in a highly concurrent VM farm, I recommend leveraging three resource pools for SQL Server-on-VMware environments. Tier-1 can be created with a high value of resources assigned for CPU and memory; Tier-2 is normal; Tier-3 is low. Do not manually specify the amount of shares for each, as this metric will become skewed if compute hardware is added or removed from the host cluster.

Read on to understand why and how, as well as a few more tips around resource pools.

Comments closed

Dynamic Resource Scheduling and SQL Server

Published 2020-03-17 by Kevin Feasel

David Klee offers some advice on Dynamic Resource Scheduling:

Dynamic Resource Scheduling (DRS) should be enabled for all VMware host clusters, especially those who run SQL Server. It provides for resource consumption load balancing functionality into a host cluster. Consider enabling the DRS load balance based on consumed memory rather than active memory (available as of vSphere 6.7), as the active memory counter for SQL Server-based VMs is not a true representation of memory usage by the SQL Server layer.

Read on for more details.

Comments closed

Improving vCenter Performance Metric Logging

Published 2020-03-10 by Kevin Feasel

David Klee has some recommendations on settings for vCenter performance metric collection:

The default performance metric statistics collection interval within vCenter is to start rolling up data in an aggregation method starting at just one hour. Much of the data necessary for troubleshooting of performance challenges reported either same day or on the previously day is lost from the vCenter data and forces the administrator to revert to cumbersome and/or time-consuming tooling, such as vRealize Operations Manager. DBAs might not have access to such tools. Hopefully by now they have read-only access to vCenter!

The vCenter performance statistics collection and rollup settings can be customized to provide a longer window of time for critical metrics to be available to the administrator for management.

Click through for some recommendations of aggregation intervals and collection durations to help with virtual machine troubleshooting.

Comments closed

Setting Up a SQL Server Lab with AutomatedLab

Published 2020-03-04 by Kevin Feasel

Jess Pomfret looks at a very interesting Powershell module:

There is a fantastic PowerShell module called AutomatedLab that can enable you to easily build out a lab for the  specific scenario you need to test. Even better is the module comes with 70 sample scripts that you can start with and adapt to meet your needs.

The module gives you the option to work with Hyper-V or VMWare. I will say most of the examples are using Hyper-V, and that is what I’ll be using also.

For my lab I want a SQL Server 2019 instance joined to a domain, and a separate client machine that I can manage the SQL Server from. On the client I would need to be able to connect to the internet as I want to be able to download PowerShell modules from the gallery easily.

It’s about time for me to rebuild my lab, so I’ll need to check that out.

Comments closed

The Costs of Virtualization

Published 2020-01-08 by Kevin Feasel

David Klee points out that virtualization, configured correctly, should not harm SQL Server performance much:

A wonderful reader of my blog sent me a note (thanks Jess!) about a single line notation in the latest SQL Server release notes. The notes is as follows.

Running SQL Server on a virtual machine will be slower than running natively because of the overhead of virtualization.

The question was simple. Why would Microsoft add this disclaimer? It was being used as a negative talking point towards SQL Server virtualization, and holding the DBA team back from getting the benefits of virtualization.

David gives us some rough numbers on what that means. Spoiler alert: if you set up your environment right, it’s not much.

Comments closed

Building a SQL Cluster Lab

Published 2019-10-31 by Kevin Feasel

Ryan Adams has started a series on building a Windows cluster in Hyper-V and layering SQL Server on top of it:

Before we start to build a SQL Cluster Lab, let’s look at the desired result. You will build a 3-node cluster replicating an environment that has two data centers. As a result, the first two nodes will reside in data center 1 and the third node in data center 2. We are creating this architecture because it is the most common architecture I see for Availability Groups. It is multi-subnet and can solve for both HA and DR.

You will notice the domain controller in the middle. That piece is certainly not representative of a production environment. However, we are using it in our lab for several different functions and being a router is one of them.

Part 1 is the only part which is currently up, but this looks like it will be a good one. Go buy a couple more sticks of RAM for your PC and get reading.

Comments closed

Understanding Power BI Data Virtualization Queries

Published 2019-08-30 by Kevin Feasel

Gerhard Brueckl walks us through a few examples of queries Power BI makes when virtualizing data:

Even though this query only touches two different data sources, it is a good way to analyze the queries sent to the data sources. To track these queries I used the built-in Performance Analyzer of Power BI desktop which can be enabled on the “View”-tab. It gives you detailed information about the performance of the report including the actual SQL queries (under “Direct query”) which were executed on the data sources. The plain text queries can also be copied using the “Copy queries” link at the bottom.

Read on for the queries and for Gerhard’s analysis.

Comments closed

Beware CPU Oversubscription with SQL Server

Published 2019-03-08 by Kevin Feasel

Monica Rathbun shares a tale of terror:

Recently I had a client complain of chronic high CPU utilization. The performance of their SQL Server had degraded, and it appeared to be related to higher than normal CPU utilization in conjunction with symptoms of unresponsive user queries.  The root cause was twofold—a third party hosting provider had overallocated virtual processors on the physical host where the virtual machine (VM) running SQL Server was residing, as well as a recent upgrade from a version of VMWare that was not patched for Spectre and Meltdown. The host had 16 physical cores and was hyperthreading (making it effectively 32 cores) until the hosting provider patched from VMWare 5.5 to a newer release (we believe 6.5) which was required for Meltdown and Spectre processor vulnerabilities. This patch disabled hyperthreading from the hypervisor to mitigate the security risk from speculative execution. Note, this patch is over a year old and a critical security risk; most software vendors (including VMWare) put this out as an immediate requirement after the announcement of the vulnerabilities.

Given this was a virtual machine, it shared a physical host with many other VMs; this is a very common configuration. However, this host was VERY overallocated.  As mentioned above, there were 16 cores–however 61 additional vCPUs had been allocated to other machines. That’s 4.3 times the number of CPUs available for allocation.  The screenshot below shows this singular Host, highlighting the vCPUs allocated.

So, uh, that’s a bad thing. Monica explains in detail why exactly it’s a bad thing, which is helpful when you’re trying to explain to the server admin why it’s a bad thing. CPU oversubscription can work for things like dev boxes or web servers, where they typically aren’t anywhere near 100% utilization. It does not work at all for busy database servers.

Comments closed