Running Redis in a Master-Slave Replication Setup

This post explains how to setup Redis distributed cache in a simple master-slave configuration, where replication is automatically is taken care of by Redis.

Key Takeaway:

Redis has the facility to setup replication with the help of master-slave configurations. A master can have multiple slaves and a slave can further have slaves. In this article we will focus on a simple setup having a single master and two slaves and will discuss a general usage pattern which would allow for a robust setup.

Read on:

Redis allows for configuration in two flavors:

1. With the help of a configuration file,

2. At runtime through commands.

In this article we will setup both the master and slaves with the help of configuration files, as that is something more understandable and how instances are configured in a production environment.

Download the Redis for Windows from MSOpenTech’s GitHub Release page for 64 bit architecture, or if you are having a 32 bit computer, please refer to my previous article on how to compile Redis for a 32 bit Windows environment. Put the folder having all the files needed to run Redis in C drive.

Next we are going to discuss some of the settings required to be implemented in the configuration files, one for each instance of Redis. The general architecture that I am trying to produce here looks something like the following:

Master-Slave-Client Diagram






























Figure 1 Redis replication basic architecture

Explanation of the architecture:

In Redis, master instance allows for both reads and writes, and in addition to that allows for disk persistence. Slaves, by default, are read only and allow for disk persistence. Over here, since this is just an introductory article, we are going to learn how to setup the simplest master-slave configuration. A more prudent setup would allow master to engage only with memory writes, and offload disk persistence to one of the slaves, and one or more slaves will dedicatedly handle the read queries. But we will discuss this in some later article.

In order to implement the aforementioned architecture we need to create three configuration files, one for master and one for each of the two slaves.

1. Nomenclature of configuration file:

It is important to name a Redis configuration file in such a way that the purpose and some vital information contained can be gleaned off from the name itself.

We will follow the pattern: redis.hosting environment.type of instance.purpose of instance.port number.conf

So a configuration meant for a master instance would bear the name like redis.windows.master.writes.5555.conf.

2. Creation of configuration files:

Redis master: Copy the configuration file that comes pre-packaged with Redis and rename it to redis.windows.master.writes.5555.conf, where 5555 is the port that will be dedicated to master instance. You can name it differently according to the port availability on your machine. Open the configuration in a text editor and change the default port from 6379 to the one that is available in your machine.

00 Master Configuration

Figure 2 Configure master instance to run on port 5555

Redis slave 1: Make a copy of the master’s configuration file and name it like redis.windows.slaveof5555.reads.5556.conf. Change the port in the file to 5556 or something that is available on your machine. Now search for the “Replication” section and un-comment the setting of “slaveof” and provide the IP address on which the master instance will be hosted and the port number. Since we will be just running all the three instances locally, the IP address should be 127.0.0.1 and the port number used in the master’s configuration file. The slave instance that we will run will take it’s configuration from this file.

00 Slave Configuration

Figure 3 Configure slave instance to receive synchronization from master

Redis slave 2: Repeat the aforementioned steps, with the exception of changing the port number to 5557 or something that is available on your machine and accordingly use the same port in the name of the file. I have named mine to redis.windows.slaveof5555.reads.5557.conf.

3. Running instances

Redis master: Open a command prompt and navigate to the folder where you are having Redis executable files and execute redis-server.exe in conjunction with the name of the configuration where fro it is supposed to pick it’s configuration from.

01 Redis Master

Figure 4 Master instance receives requests for data sync from slaves

Redis slave 1: Open another command prompt and again run the redis-server.exe file, this time specifying the slave configuration file. This will enable running a slave instance connected to the master. As soon as the slave instance will come up, the master will receive a request from slave for synchronization of data. This is shown in the screenshot.

02 Slave 1

Figure 5 Slave 1 receives and syncs data with master

Redis slave 2: Repeat the aforementioned step for slave 1, but with the other configuration file meant for slave 2.

03 Slave 2

Figure 6 Slave 2

Now run another command prompt and run redis-cli.exe and connect to the master instance. Insert some keys in the master and query them, just to make sure, they have gotten stored. Now disconnect from the master instance and connect to the first slave hosted on port 5556 (or where you hosted it.) and query for the same keys, that you inserted in the master. You will find them. Similarly you will find the same information synchronized in slave 2.

04 Redis Client

Figure 7 Redis client shows that master and slaves are at parity

Conclusion:

Running replication in Redis is very simple and minimal configuration. The pattern shown here, is elementary, just to give an idea about Redis replication. There are more robust architectures that should be used in production settings.

Running Redis as a Service in 32 bit Windows

18. January 2016 06:57 by Parakh in 32 bit, Redis, Distributed Cache, Windows  //  Tags: , , , ,   //   Comments
This article shows how to run Redis as a service in a 32 bit Windows environment.

Key Takeaway:

Learning a lot of NoSQL data-stores is easy in part because all you have to do in order to run the server, is to click on an executable file and it starts listening on a local IP address and port. Redis works the same way. But in order to run Redis in production environment, one cannot rely on a console application listening on a port. After all, there’s always a risk of someone closing the console application and shutting down the entire cache.

To remediate this, we run such products as services. Redis is no exception and in this article we will learn how to run Redis as a Windows service.

Read on:

In my last article I showed how to compile and run Redis in a 32 bit Windows environment. In this article, I am going to use the same build to run it as a Windows service.

NOTE: If you have a 64 bit processor computer, you can directly go to this link and download the binaries, in ready to use condition.

NOTE: It is always preferable to run Redis on non-default port, as that gives us the option of running an ad-hoc instance of Redis server for quick experimentation and learning. In this exercise we are going to run Redis as a service in 32 bit Windows environment and provide the runtime configuration via a configuration file.

1. Navigate to the folder with a sample configuration file. In the source code that folder will be

redis-3.0->msvs->setups->documentation->redis.windows.conf.

2. Locate the line that signifies the port on which the Redis will listen for connections. This is found in “General” category. To search, use the word “port” and change the port number to something that is available on your machine. For this exercise, change it to 6377.

Redis configuration

 

 

 

 

 

 

 

 

 

 

 

3. Copy the file in the same folder as redis-server.exe. If you compiled the code that folder will be located at the following location: redis-3.0->msvs->Win32(*mine is 32 bit environment)->Release(* compile configuration). For more information on how compile Redis source code for 32 bit Windows environment, refer to my last post.

4. Open a command prompt and navigate to the location where you are storing the Redis executable named redis-server.exe and execute the following commands:

a. C:/Redis-3.0>Redis-server - -service-install redis.windows.conf  - -service-name “Redis Server”

b. C:/Redis-3.0>net start “Redis Server”

c. C:/Redis-3.0>Redis-cli –h 127.0.0.1 –p 6377

Test Redis installation as a service

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5. If the service installation was successful and it started successfully, then you will be able to ping the server and receive a response back.

6. Just in case you decide to uninstall the service, issue the following commands:

a. C:/Redis-3.0\msvs\Win32\Release>net stop “Redis Server”

b. C:/Redis-3.0\msvs\Win32\Release>sc delete “Redis Server”

Stop and uninstall Redis service

 

 

 

 

 

 

 

 

 

 

Now you can go ahead and practice with Redis command line interface by connecting to this service.

Running Redis in 32 bit Windows

1. January 2016 00:45 by Parakh in NoSQL, Redis, Distributed Cache, Visual Studio, 32 bit  //  Tags: , , , ,   //   Comments
In this article we discuss how to compile Redis code released for Windows in Visual Studio to run it in a 32 bit Windows environment.

Key Takeaway:

Redis started its life in Linux environment and to this date, it is officially supported only in Linux environment. However, the good folks at Microsoft had started a new division with the sole purpose of porting such useful products to Windows environment and let developers use them with confidence and with the support of Microsoft behind them. Even though that division has now merged back into Microsoft, engineers at Microsoft continue to make contributions to ported projects.

With the 64 bit architecture becoming the de-facto standard in commodity computers, open source contributors and organizations around the world have started to focus on releasing binaries for 64 bit architectures. That leaves people like me who are living perfectly fine life with their old trusty computers having 32 bit processors powering them. The silver lining in this scenario, is that if you have access to the source code of a product, then you can compile it for 32 bit architecture and then use it.

This article shows how to download Redis source code, compile it using Visual Studio for a 32 bit architecture and then run it in a Windows environment.

Read on:

If you are a .Net developer, chances are that you have for majority of your professional life, written code on Windows platform, for Windows platform. On top of that if you have an old computer powered by a 32 bit processor, then you are striving to find ways to learn the new and up-coming technologies, such as NoSQL databases – MongoDB, Redis etc. which primarily have been released for 64 bit architectures.

I have been working with Redis since the past couple of months now and started learning Redis on my personal laptop which at the time of writing this article is 8 years young. This sometimes leaves me in a bit of a lurch as the newer projects and products are mainly focusing on releasing for 64 bit architectures. Well, all is not lost if you are willing to put up a little fight and compile the source code yourself, if it is available.

Thankfully Microsoft recognized that .Net developers should not get left behind when it comes to awesome products like Redis, just because they are not available to be run in the Windows environment, and hence they created a dedicated entity, Microsoft Open Tech Group devoted to porting these technologies in the Windows environment, and further the collaboration with the open source community. Redis is one of the projects that the group is handling at the moment.

Alright, so let’s run Redis on Windows in 32 bit architecture. Pre-requisite for accomplishing this is the newest version of Visual Studio 2013 with update 5, as without update 5, the process might not work successfully. If you do not have a paid version of Visual Studio, Community Version which is available for free here will also work.

1. Head over to Microsoft Open Tech Group’s Redis GitHub page for version 3.0 and download the source code available as a zip file.

2. Unzip the code in a folder in C drive. Make sure that the folder name in which the source code is housed bears a name with no space or special character.

3. Open the solution located in the msvs folder in Visual Studio.

4. Open the solution’s properties and go to the configuration manager and change the “Active solution configuration” to Release and “Active solution platform” to x86.

clip_image002

 

 

 

 

 

 

 

 

 

 

 

5. Now build the solution. Note there will be a few warnings that’ll come up, but ignore them.

6. Once the build completes, you will notice that there will be a new folder named “Win32” that would have gotten created. Inside this folder will be another folder “Release”. Release folder contains the final build and the executable files that we can use to run and learn Redis.

7. Locate the file “redis-server.exe” and execute it. It should come up looking something like shown here. This is the Redis server which by default listens on IP address 127.0.0.1 and port 6379.

clip_image004

 

 

 

 

 

 

 

 

 

 

8. Locate the file “redis-cli” and execute it. It should come up looking something like shown here. This is Redis command line interface and by default it sends commands on the address and port of 127.0.0.1 and 6379 respectively.

clip_image006

 

 

 

 

 

 

 

 

 

 

Lo and behold, we have Redis running in 32 bit Windows environment. Thanks to the brilliant folks at Microsoft for porting this valuable piece of technology to Windows environment.

Hello Redis

25. December 2015 12:55 by Parakh in Redis, NoSQL, Distributed Cache  //  Tags: , , ,   //   Comments
Knowledge about distributed cache systems now is fast becoming an essential skill for a developer. Redis is at the forefront of this product line.

2015 is coming to an end, and I have not written as much as I should have. It has been a busy year for me on the learning end and I am glad about it. Slowly, but surely, I am moving my career in the direction that I always wanted it to go. This year has been the year of NoSQL databases for me.

There’s a storm going on in the NoSQL database world, each one vying to grab mindshare and occupy a place in your development stack. Now that developers have started to understand the segment better than before, these data-stores can be safely classified into broad categories per their use case fit, such as:

1. Want to write everything to file in a schema-less environment – go for document oriented databases like MongoDB, RavenDB etc.

2. Want to deal with deep hierarchical data and process it real fast – go for graph databases such as Neo4J, OrientDB etc.

3. Want to store everything in-memory for fast retrieval – go for in-memory data-stores like Memcached, Redis, Hazelcast etc.

and so on.

It is the first and the third categories in the aforementioned list, that can be leveraged in general purpose applications and either are already in enterprise-ready state or will be in a release or two.

I recently started learning about distributed caching solutions for one of my open source projects and thus began my journey to learn one.

A distributed cache is built upon the fundamental idea of separating out the caching component from the ones provided by programming frameworks like .Net and be hosted independently. When this idea spans several servers, it becomes distributed in nature, hence the name distributed cache.

Some of the major distributed cache products available can be found over at Wikipedia. Of particular interest are Memcached, Riak, Redis and Hazelcast, as they are free for any use and have got a wide community support. At the core of all products is the central idea of storing data in RAM as a key-value pair. Of course, different products differentiate from one another on the basis of features.

It has been a short while since I have been working with Redis, and I have decided to include it in all my web projects, if I have my way.

While this article is not having anything usable technical stuff or practical utility for that matter, I just wanted to write something, and at this time, Redis is all I have in my mind. I hope the departing year was a fruitful one for you, and I wish my best for the upcoming year.

Maintaining deep hierarchical data in SQL Server

5. April 2015 12:19 by Parakh in SQL Server  //  Tags: , ,   //   Comments

Key Takeaway:

Every development project in some way has to deal with hierarchical data. That comes in the form of catalogs, cascading choices etc. In my previous article I demonstrated a way to maintain simple 2 level hierarchical data. In this article I will demonstrated the most optimal way of storing deep hierarchical data, that cannot be efficiently maintained using key-value pair setup. The way to do this is by using a dedicated data type in SQL Server – HierarchyId.

The article will not go deep into explaining about hierarchyid or the methods that accompany it, but rather will show how to use the data type in real world scenarios.

Read On:

When you have to store a complex hierarchy, traversing and manipulating that hierarchy becomes a tedious task. Nested hierarchy that goes beyond a flat 2 tier level is better managed with the help of a dedicated data type in SQL Server – HierarchyId.

HierarchyId is an in-built data-type in SQL Server and is based on CLR representation of binary data. It essentially stores an ordered path in an efficient manner. Because it is a CLR based data-types, there are methods that are available that can be leveraged in SQL queries to work with the data type. Although it is based on CLR, you do not have to take any additional steps in SQL Server to enable the use of the data type. It was first introduced in SQL Server 2008.

I will use some of the methods in this article to show you how to maintain a simple hierarchy. I will be re-using the example from my previous post and will try to store a hierarchy of different types of vehicles depending upon the environment in which they can be used.

Hierarchy of vehicles

 

 

 

 

 

 

 

 

Figure 1 Hierarchy of vehicles

In order to accommodate such a hierarchy, let’s create a data model. The data model in our case will only consists of a single table. This table can then be connected to the rest of the tables in the main data model with the help of primary-foreign key relationship.

image

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2 Table that will capture the hierarchy. Bold entries are required.

HierarchyData Table:

Column Name

Data type

Description

Id

Int

The auto incrementing integer type primary key.

Hierarchy

Hierarchyid

Binary representation of ordered path using data type hierarchyid.

HierarchyPath

Computed

A computed column to show path of a node in the hierarchy as a string for our easy comprehension.

HierarchyLevel

Computed

A computed column to show level of a node in the hierarchy as am integer value for our easy comprehension.

Entry

Navarchar(100)

Entry denoting either a parent or child.

Description

Nvarchar(1000)

Description about entry.

ShortName

Nvarchar(20)

Short string about node which will be leveraged in where clause in SQL queries.

IsActive

Bit

Bit field denoting if an entry is active or not.

CreatedBy

Nvarchar(100)

Entry to denote who actually created the entry.

CreatedDate

Datetime

Entry denoting the date and time when the entry was created.

ModifiedBy

Nvarchar(100)

Entry to denote who modified the entry.

ModifiedDate

Datetime

Entry denoting the date and time when the entry was modified.

Contrast this with the data model as presented in the previous article. You will notice that in this case we are only using a single table to store both the parent and the child entries.

As is done in the previous article, for our ease of querying, we will maintain a column called “ShortName” that will be used in where clause in all the queries. This makes the queries independent of the primary key column, in which the value may change as we move the data from one environment to another.

Now here’s the SQL that you can use to create the table and fill it up with data.

-- Creation of table
Create Table dbo.HierarchyData
(
    Id int identity(1,1) primary key,
    Hierarchy hierarchyid not null,
    HierarchyPath as Hierarchy.ToString(),
    HierarychyLevel as Hierarchy.GetLevel(),
    Entry nvarchar(100) not null,
    Description nvarchar(1000),
    ShortName nvarchar(20) not null,
    IsActive bit not null,
    CreatedBy nvarchar(100) not null,
    CreatedDate datetime not null,
    ModifiedBy nvarchar(100) null,
    ModifiedDate datetime null
);
 
-- Insertion of data
Declare 
@Root hierarchyid,
@Parent hierarchyid,
@SiblingOnRight hierarchyid,
@SiblingOnLeft hierarchyid,
@HierarchyValue hierarchyid;
 
Set @Root = hierarchyid::GetRoot();
Insert into dbo.HierarchyData Values
(@Root,'Root Level','Root Level','ROOT', 1, 'Parakh Singhal',GetUtcDate(),null,null);
 
Select @HierarchyValue = @Root.GetDescendant(null,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Ground Vehicles','Ground Vehicles','GRND_VEH',1,'Parakh Singhal', GetUtcDate(),null,null);
 
Select @SiblingOnLeft = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH';
Set @HierarchyValue = @Root.GetDescendant(@SiblingOnLeft,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Air Vehicles','Air Vehicles','AIR_VEH',1,'Parakh Singhal', GetUtcDate(),null,null);
 
Select @SiblingOnLeft = Hierarchy from dbo.HierarchyData where ShortName = 'AIR_VEH';
Set @HierarchyValue = @Root.GetDescendant(@SiblingOnLeft,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Water Vehicles','Water Vehicles','WATER_VEH',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @HierarchyValue = Hierarchy.GetDescendant(null,null) from dbo.HierarchyData where ShortName = 'GRND_VEH';
Insert into dbo.HierarchyData values
(@HierarchyValue,'Trucks','Trucks','GRND_VEH_TRUCKS',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH';
Select @SiblingOnLeft = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH_TRUCKS';
Set @HierarchyValue = @Parent.GetDescendant(@SiblingOnLeft,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Cars','Cars','GRND_VEH_CARS',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @Parent = Hierarchy.GetDescendant(null,null) from dbo.HierarchyData where ShortName = 'AIR_VEH';
Set @HierarchyValue = @Parent;
Insert into dbo.HierarchyData values
(@HierarchyValue,'Hot Air Baloon','Hot Air Baloon','AIR_VEH_BALOON',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'AIR_VEH';
Select @SiblingOnLeft = Hierarchy from dbo.HierarchyData where ShortName = 'AIR_VEH_BALOON';
Set @HierarchyValue = @Parent.GetDescendant(@SiblingOnLeft,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Airplane','Airplane','AIR_VEH_AIRPLANE',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'WATER_VEH';
Set @HierarchyValue = @Parent.GetDescendant(null,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Ship','Ship','WATER_VEH_SHIP',1,'Parakh Singhal',GetUtcDate(),null,null);
 
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'WATER_VEH';
Select @SiblingOnLeft = Hierarchy from dbo.HierarchyData where ShortName = 'WATER_VEH_SHIP';
Set @HierarchyValue = @Parent.GetDescendant(@SiblingOnLeft,null);
Insert into dbo.HierarchyData values
(@HierarchyValue,'Submarine','Submarine','WATER_VEH_SUBMARINE',1,'Parakh Singhal',GetUtcDate(),null,null);

And here’s the SQL that you can use to query some of the most common scenarios like:

1. Traversing down the hierarchy from a parent node,

2. Traversing up the hierarchy from a child node,

3. Finding the level of a node in the hierarchy,

4. Determining is a node is the child of a node,

5. Finding the ancestor of a child node.

Declare
@Root hierarchyid,
@Parent hierarchyid,
@Child hierarchyid,
@Node hierarchyid;
 
-- Selecting the entire table contents.
Select * from dbo.HierarchyData;
 
-- Finding the root element.
Set @Root = hierarchyid::GetRoot();
Select * from dbo.HierarchyData where Hierarchy = @Root;
 
-- Climbing down a hierarchy
-- Selecting a hierarhcy of nodes belonging to a parent.
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH';
 
Select * from dbo.HierarchyData 
where Hierarchy.IsDescendantOf(@Parent) = 1;
 
-- Climbing up the hierarchy
-- Selecting all the parents of a node
Select @Node = Hierarchy from dbo.HierarchyData where ShortName = 'WATER_VEH_SUBMARINE';
 
Select * from dbo.HierarchyData 
where @Node.IsDescendantOf(Hierarchy) = 1;
 
-- Getting level of a node in the hierarchy.
Select Hierarchy.GetLevel() as HierarchyLevel from dbo.HierarchyData where ShortName = 'AIR_VEH_AIRPLANE';
 
-- Determining if a node is a child of a node.
Select @Parent = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH';
Select @Child = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH_TRUCKS'
 
Select @Child.IsDescendantOf(@Parent) as Result;
 
-- Determining an ancestor
-- The query selects the immediate ancestor.
Select @Node = Hierarchy from dbo.HierarchyData where ShortName = 'GRND_VEH_CARS';
Select * from dbo.HierarchyData where Hierarchy = @Node.GetAncestor(1);

Brief overview of some of the methods used in aforementioned sql queries:

1. GetRoot(): GetRoot methods when used on a hierarchyid type value will return the hierarchyid type value of the absolute root of the hierarchy.

2. GetDescendant(): Methods when used on a hierarchyid type value will return a hierarchyid type value. You can use this method to generate a new hierarchyid value when inserting data, the way I have demonstrated.

Notice that in order to create nodes that are siblings, you need to specify a pre-existing sibling (See the creation of water and air vehicles that are siblings to ground vehicles). Depending upon where you want to place the sibling in the hierarchy, you might need to provide both the sibling on the left and right in the hierarchy. Since I want to treat both air and water vehicles, I only mentioned sibling on the left in the hierarchy. If you want to place another category, say, amphibious vehicles between water and ground vehicles, then you will have to mention ground vehicles as the sibling on the left and water vehicles as the sibling on the right.

3. IsDescendantOf(): Method takes in a hierarchy id as input and returns a Boolean, if that is true.

4. GetAncestor(): Method takes in an integer value and gives the parent of the child node, depending on the integer value provided. The integer depicts the level in the hierarchy.

As you can see in the code above, maintaining a deep hierarchy with the help of hierarchyid is easy. It does not require cumbersome recursive Common Table Expressions and multiple tables to maintain, and yet offers more flexibility.

I have not gone deep into explaining about hierarchyid and the methods that accompany it. The main motive of this article is to demonstrate with the help of working code, how to maintain hierarchical data with the help of in-built mechanism in SQL Server.

References:

1. Model Your Data Hierarchies With SQL Server 2008

2. SQL Server 2008 - HierarchyID - Part I

3. Hierarchies with HierarchyID in SQL 2008

Month List

ParakhSinghal@Twitter.com

Note: For Customization and Configuration, CheckOut Recent Tweets Documentation