Press "Enter" to skip to content

Category: Cloud

The Tradeoffs of Azure SQL Managed Instance General Purpose

Published 2024-01-23 by Kevin Feasel

Kendra Little provides some guidance:

Whether or not you use Azure SQL Managed Instance, you will likely be asked for an opinion on it eventually if you’re a SQL Server person.

While the architecture is documented, it can be a bit of a long read and some of the gotchas are spread out over different pages – so I’m drawing up the architecture of each service tier along with notable implications for the design on performance and cost. Here’s the scoop on General Purpose.

Click through for Kendra’s thoughts.

Comments closed

Tips for Saving Money in the Cloud

Published 2024-01-22 by Kevin Feasel

David Klee offers up some advice:

You might be able to shave off thousands – or more – in your monthly cloud bills for your critical SQL Servers, all while maintaining or even improving performance.

Public cloud providers charge organizations for everything they deploy, and while a few items in the cloud are based on a pay-by-consumption model, most of the services that really stack up, namely SQL Server licensing, are charged on a pay-by-allocation model. If you provision a SQL Server VM with eight cores and promptly forget about it, you will be shocked at the end of the month when you receive your cloud subscription bill. If you provision a 32-core VM, and your workload only uses four cores, you are still paying for all 32 cores, regardless of the low utilization rates. The same goes for memory and storage. A large memory footprint and lots of high-speed attached managed disks all add up.

Read on for some advice around figuring out the right server size for your workload.

Comments closed

Exporting to CSV in Azure ML Designer

Published 2024-01-19 by Kevin Feasel

Tom LaRock saves a file:

The most popular feature in any application is an easy-to-find button saying “Export to CSV.” If this button is not visibly available, a simple right-click of your mouse should present such an option. You really should not be forced to spend any additional time on this Earth looking for a way to export your data to a CSV file.

Well, in Azure ML Studio, exporting to a CSV file should be simple, but is not, unless you already know what you are doing and where to look. I was reminded of this recently, and decided to write a quick post in case a person new to ML Studio was wondering how to export data to a CSV file.

Click through for one false start and then the correct answer.

Comments closed

Storing SQL Server Backups in Cloudflare R2

Published 2024-01-19 by Kevin Feasel

Daniel Hutmacher saves a buck:

R2 is Cloudflare’s own implementation of AWS S3 storage, with some big benefits – one of them being no egress fees, which is great if you want to publish or distribute a lot of data (like I did with this demo database). In this post, I thought I’d briefly document how to set up R2, and how to use it to back up and restore your SQL Server databases.

You’ll need a Cloudflare account to follow along. The account and a lot of their services are free, but R2 storage obviously comes with a small cost. For scale, I’m running an almost-terabyte bucket at just a couple of dollars per month.

Given the number of times I’ve pushed Daniel’s excellent Chicago parking tickets database (including right now—it’s a great database that I’ve used in several presentations and videos!), the lack of egress charges is pretty big.

Comments closed

Sample Data in Azure ML Designer

Published 2024-01-15 by Kevin Feasel

Tom LaRock shows us where the hidden data is:

Recently I was working inside of Azure ML Studio and wanted to browse the sample datasets provided. Except I could not find them. I *knew* they existed, having used them previously, but could not remember if that was in the original ML Studio (classic) or not.

After some trial and error, I found them and decided to write this post in case anyone else is wondering where to find the sample datasets. You’re welcome, future Tom!

Click through to see where those sample datasets are. And yeah, they don’t get updated that frequently. And that’s probably a good thing, as it means when you run the demo two years after someone created it, you’ll still get predictable results.

Comments closed

Comparing Fabric F2 to F64

Published 2024-01-11 by Kevin Feasel

Reitse Eskens enters austerity mode:

If you’ve been having fun with Microsoft Fabric, chances are you’ve been playing around with the F64 capacity trial. This one is given to you by Microsoft for free but, since the GA data, the timer attached to it is counting down the days until you need to buy your own.

Read on to see what happens when you lose out on that sweet F64 goodness. I actually do appreciate the way that Fabric works: it’s not a linear scale of “F2 means you get 1/32 the processing power of F64.” Rather, it’s closer to time slices on a mainframe: F64 gets you a bigger slice. So if you’re a small shop without an enormous amount of data, F2 really does work pretty well.

Comments closed

Goodbye Aurora Serverless v1

Published 2024-01-09 by Kevin Feasel

Alex Woodie breaks the news:

AWS has notified customers of its Amazon Aurora Serverless v1 service that it will cease supporting the offering at the end of 2024. Replacing v1 in the Aurora Serverless range, which supports Postgres and MySQL databases, will be v2, which offers some advantages but also one big disadvantage: It doesn’t scale all the way down to zero.

Click through for more information.

Comments closed

Optimized Locking in Azure SQL DB

Published 2024-01-09 by Kevin Feasel

Aaron Bertrand tries out a new feature:

In a sentence: Instead of locking individual rows and pages for the life of the transaction, a single lock is held at the transaction level, and row and lock pages are taken and released as needed.

This is made possible by previous investments in Accelerated Database Recovery and its persistent version store. A modification can evaluate the predicate against the latest committed version, bypassing the need for a lock until it is ready to update (this is called lock after qualification, or LAQ). There’s a lot more to it than that, and I’m not going to dive deep today, but the result is simple: long-running transactions will lead to fewer lock escalations and will do a lot less standing in the way of the rest of your workload. Locks held for shorter periods of time will naturally help reduce blocking, update conflicts, and deadlocks. And with fewer locks being held at any given time, this will help improve concurrency and reduce overall lock memory.

Read on to learn more about how it works and Aaron’s initial thoughts on the feature.

Comments closed

Managing Azure Data Factory IP Ranges for Azure Firewalls

Published 2024-01-03 by Kevin Feasel

Meagan Longoria has a script for us:

While a private endpoint and vNets are preferred, sometimes we need to configure Azure SQL Database or Azure Storage to allow use of public endpoints. In that case, an IP-based firewall is used to prevent traffic from unauthorized locations. But Azure Data Factory’s Azure Integration Runtimes do not have a single static IP. So how do we keep the firewall updated so that ADF can access these resources?

One option is to run everything through a self-hosted integration runtime, which can have a static public IP. But that also means you have to install (and keep updated) a Java SDK if you are converting data to Parquet or ORC files.

Another option is to obtain the IP range list published by Microsoft and update the firewall. That is what I did for a recent project.

Read on to learn more about the latter option, including a Powershell script to do the work.

Comments closed