Anyone that moved from server to data center successfully or in the process of doing so?

Sara February 26, 2020

Hi I am in the process of migrating from server to data center. There are several roadblocks that I had gone past and several more on my way. Anyone on the same boat? If yes, I would like to discuss online/ meet at the summit.

4 answers

Suggest an answer

Log in or Sign up to answer
1 vote
Steve Tsiopanos February 28, 2020

Hi @Sara

I just completed my company’s migration to Jira Datacenter, late last year, after planning / preparing for over a year.

The biggest obstacle, in my experience, was finding the ideal intersection of full add-on compatibility and Jira version “freshness” (Software and Service Desk). After we figured out what the technology stack should look like, which was it’s own significant challenge, it took several months before we could hit an optimal upgrade window.

It seemed like through at least late summer 2019, most add-on developers had not quite caught up with Data Center compatibility. Also, it didn’t help that Atlassian was continuously evolving its guidelines for JDC approval. I cannot fault them too much, however, due to the complexity of encouraging compatibility with JDC and Jira Cloud without alienation developers with fewer resources at their disposal. 

Overall, the migration from Software to Datacenter is fairly straightforward, essentially boiling down to a license change and minor configuration file change. Well... at least as far as Jira is concerned. We did have to ensure the requisite data was moved to the shared volume mount, that SQL dumps / imports were prepared, internal / external DNS configuration was complete, etc. I should write a blog post about this...

When we finally moved to JDC, we immediately noticed improved performance under heaving reporting load. Our international users, in particular noticed improvements to responsiveness due to our implementation of AWS CloudFront as the CDN (love that JDC supports that natively!). 

Beyond that, all your regular Jira Administrator and SysAdmin Certification knowledge applies :) 

- Archive old tickets / projects where possible

- Reduce / consolidate / remove Custom Fields where possible and scope them down

- Generally speaking, try to trim down the amount of Schemes used, across the board. 

- Scope down Active Directory user / group synchronization where possible

- Watch your Automation Rules! Scope them down as much as possible.

- and more... Atlassian writes volumes on this stuff, I’m just scratching the surface

The Jira instance I mention here dates back to 2011 (when I first built it) so it has picked up a lot of technical debt over time. Performance optimization is a regular duty. Soon, I expect we will work with our F5 engineer to route certain types of traffic to dedicated reporting nodes so we can, for example, maintain optimal performance for Jira Service Desk users, regular Scrum board users, etc.

 

***

For anyone interested, our technology stack might be slightly on the unique side...

Infrastructure consists of:

5 app nodes
3 MySQL Galera Cluster nodes
3 GlusterFS nodes

OS - CentOS 7 

Deployed into a vSphere vm hosting cluster.

Load-balancing and TLS offloading is provided by a physical F5 cluster pair.

We use the GlusterFS cluster to provision a brick of distributed HA storage which is mounted to the Jira shared directory on each app node (for attachments and other shared data).

The MySQL Galera cluster works surprisingly well, giving us a 3 node replica set with automatic self-healing synchronization. We keep it distributed across ESXi cluster hosts (in our data center). 

The GlusterFS cluster, being distributed across discreet virtual machines, allows us to fully maintenance any single node with zero-downtime for the shared volume.

Note: On each node I implemented HAproxy to proxy connection requests from port 443 (plus self-signed cert) to port 8080. This means we can configure end-to-end encryption from the F5 load balancer directly to each Jira DC node, while maintaining the Certificate Authority signed cert on the F5 (where it’s easy to change it without affecting application uptime). 



Peter T
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
February 28, 2020

Hi @Steve Tsiopanos , amazing story. Are you interested to write a blog post about it?

Steve Tsiopanos February 28, 2020

Thanks @Peter T . Time permitting, I would love to write a blog post about our experience. Hopefully I can publishing something before Atlassian Summit.

Like Sara likes this
Peter T
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
February 28, 2020

Excellent, if you want, we will be happy to publish it at botronsoft.com along with our other Jira Administration topics.

Cheers

Like Steve Tsiopanos likes this
Sara March 2, 2020

Hi Steve, Great to hear that your migration is completed. I am not familiar with MySQL Galera. What I need to do is data migration from on-prem MySQL into AWS Postgress RDS. My JIRA instance is constantly cleaned up similar to what you described. Hence artifacts in JIRA are controlled. What is the process of data migration here? What was your  JIRA / Confluence data size during migration?

I am having trouble with 9 GB Confluence data to be migrated - this is just the db size. attachments are moved separately.  We tried multiple migration tools such as pgloader and ESF migration tool kit. After migration, Confluence complains about data type, case and column name mismatches. This ends up in an endless process of trying to correct discrepancies as application throws error and not getting into a reliable instance. Atlassian does not support third party tools migration that is the primary reason I have been seeking Atlassian support on data migration using the built in export import tool. I only 9 GB data, and the tool is said to have worked successfully for up to 10 GB data. For the past several weeks I am working with limited help from Atlassian to fine-tune memory settings hoping that the tool will work. With all the resource fine-tuning, the resultant file size remains the same. We are running out of time. Hence would appreciate advice from any one having insight on successful use of Atlassian built-in tool for data migration.

0 votes
Gilbert Reid August 8, 2021

it was really an amazing story but i am still facing an issue on my server that is related to kbc head office no. it showing an unknow error. can you please let me know what is the best solution?

0 votes
Teodora _Old Street Solutions_Tempo_
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
February 27, 2020

Hey @Sara 

One of our solutions is based on the use case of migrating data from server to server/data center fully automatic and with 0 data loss. We also have a services team that is handling the more complex use cases, so (I dare to say) we are experts in this area.

Join us on our booth #1102 to share more about your use case; our team is always ready to support. :)

0 votes
Benjamin Zaczek February 26, 2020

Hi Sara, we're about to go on this journey ourselves over the next few months, though we are currently looking to not stand up our new DC for a few weeks. I'd love to hear inputs from others or yourself! We are merging two server systems into our DC new DC system too.

Peter T
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
February 27, 2020

@Sara @Benjamin Zaczek the jorney from multiple servers to DC has 3 main components - organizational, technical and people. This video is from last Summit at Vegas and the gentlemen from J&J talks about all these aspects. Very useful.

A Journey to Enterprise Agility: Migrating 15 Atlassian Instances to Data Center 

Let us know how do you like it.

Cheers,
Peter T

Like Benjamin Zaczek likes this
TAGS
AUG Leaders

Atlassian Community Events