Oracle GoldenGate Installation

Installing Oracle GoldenGate 23ai is more than just running a setup file—it’s about preparing your system for efficient, reliable data replication. In this episode, Lois Houston and Nikita welcome back Nick Wagner to break down system requirements, storage considerations, and best practices for installing GoldenGate.

 

You’ll learn how to optimize disk space, manage trail files, and configure network settings to ensure a smooth installation.

 

Oracle GoldenGate 23ai: Fundamentals: https://mylearn.oracle.com/ou/course/oracle-goldengate-23ai-fundamentals/145884/237273

Oracle University Learning Community: https://education.oracle.com/ou-community

LinkedIn: https://www.linkedin.com/showcase/oracle-university/

X: https://x.com/Oracle_Edu

 

Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.

 

-------------------------------------------------------------

 

Episode Transcript:

00:00

Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we’ll bring you foundational training on the most popular Oracle technologies. Let’s get started!

00:25

Nikita: Hello and welcome to Oracle University Podcast! I’m Nikita Abraham, Team Lead of Editorial Services with Oracle University, and I’m joined by Lois Houston, Director of Innovation Programs. 

Lois: Hi there! Last week, we took a close look at the security strategies of Oracle GoldenGate 23ai. In this episode, we’ll discuss all aspects of installing GoldenGate.

00:48

Nikita: That’s right, Lois. And back with us today is Nick Wagner, Senior Director of Product Management for GoldenGate at Oracle. Hi Nick! I’m going to get straight into it. What are the system requirements for a typical GoldenGate installation?

Nick: As far as system requirements, we're going to split that into two sections. We've got an operating system requirements and a storage requirements. So with memory and disk, and I know that this isn't the answer you want, but the answer is that it varies.
With GoldenGate, the amount of CPU usage that is required depends on the number of extracts and replicats. It also depends on the number of threads that you're going to be using for those replicats. Same thing with RAM and disk usage. That's going to vary on the transaction sizes and the number of long running transactions.

01:35

Lois: And how does the recovery process in GoldenGate impact system resources? 

Nick: You've got two things that help the extract recovery. You've got the bonded recovery that will store transactions over a certain length of time to disk. It also has a cache manager setting that determines what gets written to disk as part of open transactions. It's not just the simple answer as, oh, it needs this much space.

GoldenGate also needs to store trail files for the data that it's moving across. So if there's network latency, or if you expect a certain network outage, or you have certain SLAs for the target database that may not be met, you need to make sure that GoldenGate has enough room to store its trail files as it's writing them. The good news about all this is that you can track it. You can use parameters to set them. And we do have some metrics that we'll provide to you on how to size these environments.

So a couple of things on the disk usage. The actual installation of GoldenGate is about 1 to 1.5 gig in size, depending on which version of GoldenGate you're going to be using and what database. The trail files themselves, they default to 500 megabytes apiece.
A lot of customers keep them on disk longer than they're necessary, and so there's all sorts of purging options available in GoldenGate. But you can set up purge rules to say, hey, I want to get rid of my trail files as soon as they're not needed anymore. But you can also say, you know what? I want to keep my trail files around for x number of days, even if they're not needed. That way they can be rebuilt. I can restore from any previous point in time.

03:15

Nikita: Let’s talk a bit more about trail files. How do these files grow and what settings can users adjust to manage their storage efficiently?

Nick: The trail files grow at about 30% to 35% of the generated redo log data. So if I'm generating 100 gigabytes of redo an hour, then you can expect the trail files to be anywhere from 30 to 35 gigabytes an hour of generated data.

And this is if you're replicating everything. Again, GoldenGate's got so many different options. There's so many different ways to use it. In some cases, if you're going to a distributed applications and analytics environment, like a Databricks or a Snowflake, you might want to write more information to the trail file than what's necessary.

Maybe I want additional information, such as when this change happened, who the user was that made that change. I can add specific token data. You can also tell GoldenGate to log additional records or additional columns to the trail file that may not have been changed. So I can always say, hey, GoldenGate, replicate and store the entire before and after image of every single row change to your trail file, even if those columns didn't change.

And so there's a lot of different ways to do it there. But generally speaking, the default settings, you're looking at about 30% to 35% of the generated redo log value. System swap can fill up quickly. You do want this as a dedicated disk as well. System swap is often used for just handling of the changes, as GoldenGate flushes data from memory down to disk.

These are controlled by a couple of parameters. So because GoldenGate is only writing committed activity to the trail file, the log reader inside the database is actually giving GoldenGate not only committed activity but uncommitted activity, too. And this is so it can stay very high speed and very low latency.

05:17

Lois: So, what are the parameters?

Nick: There's a cache manager overall feature, and there's a cache directory. That directory controls where that data is actually stored, so you can specify the location of the uncommitted transactions.

You can also specify the cache size. And there's not only memory settings here, but there's also disk settings. So you can say, hey, once a cache size exceeds a certain memory usage, then start flushing to disk, which is going to be slower. This is for systems that maybe have less memory but more high-speed disk. You can optimize these parameters as necessary.

05:53

Nikita: And how does GoldenGate adjust these parameters?

Nick: For most environments, you're just going to leave them alone. They're automatically configured to look at the system memory available on that system and not use it all. And then as soon as necessary, it'll overflow to disk.

There's also intelligent settings built within these parameters and within the cache manager itself that if it starts seeing a lull in activity or your traditional OLTP type responses to actually free up the memory that it has allocated. Or if it starts seeing more activity around data warehousing type things where you're doing large transactions, it'll actually hold on to memory a little bit longer. So it kinda learns as it goes through your environment and starts replicating data.

06:37

Lois: Is there anything else you think we should talk about before we move on to installing GoldenGate? 

Nick: There's a couple additional things you need to think of with the network as well. So when you're deploying GoldenGate, you definitely want it to use the fastest network. 

GoldenGate can also use a reverse proxy, especially important with microservices. Reverse proxy, typically we recommend Nginx. And it allows you to access any of the GoldenGate microservices using a single port. 

GoldenGate also needs either host names or IP addresses to do its communication and to ensure the system is available. It does a lot of communication through TCP and IP as well as WSS. And then it also handles firewalls. So you want to make sure that the firewalls are open for ingress and egress for GoldenGate, too.

There's a couple of different privileges that GoldenGate needs when you go to install it. You'll want to make sure that GoldenGate has the ability to write to the home where you're installing it. That's kind of obvious, but we need to say it anyways.

There's a utility called oggca.sh. That's the GoldenGate Configuration Assistant that allows you to set up your first deployments and manage additional deployments. That needs permissions to write to the directories where you're going to be creating the new deployments.

The extract process needs connection and permissions to read the transaction logs or backups. This is not important for Oracle, but for non-Oracle it is. And then we also recommend a dedicated database user for the extract and replicat connections.

08:15

Are you keen to stay ahead in today's fast-paced world? We’ve got your back! Each quarter, Oracle rolls out game-changing updates to its Fusion Cloud Applications. And to make sure you’re always in the know, we offer New Features courses that give you an insider’s look at all of the latest advancements. Don't miss out! Head over to mylearn.oracle.com to get started.

08:41

Nikita: Welcome back! So Nick, how do we get started with the installation? 

Nick: So when we go to the install, the first thing you're going to do is go ahead and go to Oracle's website and download the software. Because of the way that GoldenGate works, there's only a couple moving parts. You saw the microservices. There's five or six of them. You have your extract, your replicat, your distribution service, trail files. There's not a lot of moving components.

So if something does go wrong, usually it affects multiple customers. And so it's very important that when you go to install GoldenGate, you're using the most recent bundle patch. And you can find this within My Oracle Support. It's not always available directly from OTN or from the Oracle e-delivery website. You can still get them there, but we really want people going to My Oracle Support to download the most recent version.

There's a couple of environment variables and certificates that you'll set up as well. And then you'll run the Configuration Assistant to create your deployments. 

09:44

Lois: Thanks, Nick, for taking us though the installation of GoldenGate. Because these are highly UI-driven topics, we recommend that you take a deep dive into the GoldenGate 23ai Fundamentals course, available on mylearn.oracle.com.

Nikita: In our next episode, we’ll talk about the Extract process. Until then, this is Nikita Abraham…

Lois: And Lois Houston signing off!

10:08

That’s all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We’d also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.