Press "Enter" to skip to content

Day: June 17, 2022

Automating Parallelism Decisions in Flink Batch Jobs

Lijie Wang and Zhu Zhu describe Apache Flink’s batch scheduler:

Deciding proper parallelisms of operators is not an easy work for many users. For batch jobs, a small parallelism may result in long execution time and big failover regression. While an unnecessary large parallelism may result in resource waste and more overhead cost in task deployment and network shuffling.

To decide a proper parallelism, one needs to know how much data each operator needs to process. However, It can be hard to predict data volume to be processed by a job because it can be different everyday. And it can be harder or even impossible (due to complex operators or UDFs) to predict data volume to be processed by each operator.

To solve this problem, we introduced the adaptive batch scheduler in Flink 1.15. The adaptive batch scheduler can automatically decide parallelism of an operator according to the size of its consumed datasets. 

Read on to see some of the benefits of using the adaptive batch scheduler, as well as some of the decision points it uses along the way.

Comments closed

Multidimensional Bloom Filters

The Instaclustr team talks bloom filters:

Bloom filters are space-efficient probabilistic data structures that can yield false positives but not false negatives. They were initially described by Burton Bloom in his 1970 paper  “Space/Time Trade-offs in Hash Coding with Allowable Errors“. They are used in many modern systems including the internals of the Apache® projects Cassandra®, Spark™, Hadoop®, Accumulo®, ORC™, and  Kudu™.

Multidimensional Bloom filters are data structures to search collections of Bloom filters for matches. The simplest implementation of a Multidimensional Bloom filter is a simple list that is iterated over when searching for matches. For small collections (n < 1000) this is the most efficient solution. However, when working with collections at scale other solutions can be more efficient. 

Read on to learn more, including some discussion about an implementation in Cassandra.

Comments closed

Query Store Hints in SQL Server 2022

Erik Darling has thoughts:

When you’re dealing with untouchable vendor code full of mistakes, ORM queries that God has turned away from, and other queries that for some reason can’t be tinkered with, we used to not have a lot of options.

In SQL Server 2022, Query Store gains a new super power: you can add hints to queries without intercepting the code in some other manner.

There are a couple of useful hints which won’t be available but Erik seems mostly upbeat about what is there.

Comments closed

Extracting Data from DAX Measures into CSV

Gilbert Quevauvilliers builds a process:

In this blog post I am going to demonstrate how to use the new Power Automate Flow to extract data from a DAX measure into a SharePoint CSV file.

I got this idea after reading the blog post from the Microsoft Power BI Team: Unlocking new self-service BI scenarios with ExecuteQueries support in Power Automate | Microsoft Power BI Blog | Microsoft Power BI

The great news is that this works on Power BI Pro, Premium Per User and Premium.

Read on to see how.

Comments closed

Refreshing SQL Managed Instances which Use TDE

Bradley Ball keeps the dev environment up to date:

Hello Dear Reader!  I was working with some friends lately and we needed to set up a process to refresh their Development Environment databases from Production.  Additionally, the databases are encrypted using Transparent Data Encryption, The SQL MI instances are in different regions, and the SQL MI Instances are in different subscriptions.  To duplicate the environment, in order to match our friends, we did the following setup.

Click through for a high-level overview, step-by-step guidance, and a whole lot of detail.

Comments closed

Reading the SQL Server Error Log

Lee Markum has two ways to read the SQL Server error log:

Reading the SQL Server Error Log is important when troubleshooting many issues for SQL Server. Some example issues would be errors related to Always On Availability Groups, long IO occurrence messages, and login failures.

Unfortunately, the SQL Server Error Log can be very full of information, making specific things hard to find, especially if you’re just visually scrolling the Error Log. Even if you’re recycling the Error Log each day and keeping 30 or more days of error log, on a busy system, the error log can still be quite full, even for a single day.

Click through for those techniques.

Comments closed

Searching for SQL Server Backup Locations

David Fowler can’t remember where those backups went:

Sometimes I find remembering where a particular server sends its backups to a nightmare.

You might have servers backing up to different locations, you might have different locations for individual databases and different locations for your fulls, diffs and logs. You might be trying to get your head around a customer’s set up, where the backups make no logical sense at all.

Whatever you’re up to, at some point, for some reason you’re going to need to access your backup location to get at the files.

Read on for a Powershell script which can help out with this task.

Comments closed