Instead of the single server and database you get with Atlassian Server, Data Center provides additional application nodes as well as a load balancer to help distribute traffic between nodes during peak usage so you can scale your Atlassian products and offer high availability.
The components of your Data Center infrastructure will include:
- Load Balancer
- Application Nodes
- Clustered Database
- Shared File System
While Atlassian does not provide specific recommendations for your hardware, we will share some statistics from a survey conducted by Atlassian on their customer base, later on in this post.
In the meantime, let’s take a deeper look at the components required for your Data Center migration and the other essential considerations to plan a Server to DC migration.
The 4 Essential Elements Needed for a Server to Data Center Migration
As requests come in over the network, they will be managed by the load balancer. This piece of equipment or software will then balance the incoming traffic amongst the application server nodes in your cluster to ensure you have the performance you need and the availability your users demand.
Most load balancers can handle your SSL encryption as well. This allows you to remove that load from your application servers, while ensuring that your data is kept secure while in transit.
According to an Atlassian Data Center survey, customers selected from a wide variety of load balancers including HAProxy, BrocadeADX, Apache, VMWare, CISCO, Netscaler, and WAS ELB.
Both hardware and software load balancers are supported, however, you are required to configure your chosen load balancer for cookie-based session affinity, a.k.a. sticky sessions.
Essentially, this means that when a user accesses the application, they will remain on the same node throughout their session.
These nodes are where your Atlassian application lives. Each node in your environment will have a copy of your software installed on it whether it be Jira Software, Jira Service Desk, Confluence, Bitbucket, or Crowd.
Each node in your cluster will be available to serve your users, but they must all be operating the same version of the application and they must all be clustered in the same physical location.
Most customers have between two and four nodes. We recommend starting with two or three nodes and adding additional nodes as needed.
To ensure high availability, we also recommend deploying your Data Center application on one node, to begin with. Test functionality to make sure it’s working as anticipated before introducing an additional node. Once you have two nodes, test your load balancer to ensure that it’s directing traffic as anticipated. Repeat this process for each additional node added.
For Atlassian Data Center, your database is required to be installed on its own node, separate from your application nodes.
Clustered database technology is supported in Data Center and is recommended as it is proven to be more resilient at combating downtime.
Data Center supports the same databases as a Server instance, but we recommend double-checking that your preferred database technology and version are supported before pulling the trigger on a purchase.
According to their survey, 46% of customers utilize PostgreSQL, 31% use MSSQL, 15% Oracle, and 8% MySQL. Each technology has an option to cluster and make your database highly available.
Shared File System
Your shared file system will store your Apps, attachments, icons and avatars. Like your database, the shared file system is connected to each of the application server nodes..
SAN, NFS, and NAS file sharing protocols are supported, but distributed protocols like DFS are not.
Critical DevOps Considerations for Your Data Center Instance
Regardless of your Data Center use case, it’s a good idea to have a DR system implemented to maintain availability in the event your primary site goes down. To achieve this, deploy an offsite disaster recovery system in a second location.
The DR site typically mirrors the functionality of the primary site, however, you may choose to make it less powerful to save cost. Perhaps limit to one application server node. Once the DR node is up and running, more application nodes can be added if you need to run on the DR site for an extended amount of time.
You can execute a database replication based on the database technology you’ve chosen and copy your production database into your DR system.
Finally, you’ll want to make sure to synchronize the shared file system to your DR site at roughly the same interval as your database replication. For more information on implementing your DR strategy, or help to execute replication, see the Disaster Recovery Guide for Jira Data Center.
You have many options for hosting your Data Center instance, whether it be bare metal servers, virtual machines, or a cloud hybrid hosted environments like AWS or Azure.
In the survey mentioned above, 85% of Data Center installations were at least partially virtual via the newer infrastructure as service offerings we mentioned above. This option is becoming increasingly popular, as most virtual and cloud service providers have features such as snapshotting, ease of configuration and more.
It may be tempting to geographically distribute your application server nodes, but keep in mind that Atlassian does not support this type of distribution. We recommend that all your nodes are located in the same geographical region when using an IaaS option in order to remain in a supported configuration.
Data Center for Enterprise Scale
Atlassian has made it their mission to give enterprise clients with complex needs the tools to succeed with Data Center.
To that end, they have partnered with Solution partners, like Coyote Creek, to provide proactive and strategic guidance in planning, executing, and optimizing your Atlassian investment.
At Coyote Creek, we can also assist you with all Atlassian Data Center, Cloud, or Server licenses including co-termination and upgrades.
Our senior IT consultants can help you to identify issues within your existing environment during a pre-migration audit and help you correct these issues to optimize your new Atlassian Data Center environment.
We will advise on optimal usage of advanced Data Center functions like SAML integration with Single Sign-On (SSO,) archiving, Disaster Recovery, etc. and provide a roadmap to help you protect your investment.
To learn more about how we can work closely with you to assist with your installation and configuration, contact us today.