Please see below some frequently asked questions regarding migrating from CentOS 7 to Rocky/RHEL 9 operating system.
Is there a cost for this move?
Jadu will undertake this server maintenance for all Jadu hosted customers with an active support agreement at no additional cost.
Why do we need to migrate away from CentOS 7?
CentOS 7 has been set an official end of life date during 2024, after which time we can expect no security patches from the community maintaining this project.
What is Rocky 9? Will we notice any difference?
Like CentOS 7, Rocky 9 is an open-source variant of the Red Hat family of operating systems. Being from the same family, the two are very similar in how they operate. There should be no noticeable difference when your site has been migrated over.
Why are you not moving to another version of CentOS?
The CentOS project has, historically, been ‘downstream’ of Red Hat Enterprise Linux (RHEL). This means that any fixes are made available to CentOS after test and release in RHEL, usually within 24 hours.
CentOS 9 has moved to be ‘upstream’ of RHEL. In other words, it will receive fixes ahead of RHEL, more at the bleeding edge and therefore potentially less stable. Jadu does not consider this suitable for use in our PROD environments at this time.
We have selected Rocky as this maintains the stability of being ‘downstream’ of RHEL. It is seen as a spiritual successor to CentOS, having been created by the person originally responsible for CentOS.
When will our server(s) be moved?
We will contact you over the coming weeks via Jadu Support with a plan for your specific infrastructure as part of a wider programme of work that Jadu is undertaking for all hosted customers.
Can you tell us what is involved in the process?
We will implement the process on your UAT environment first to give you an opportunity to test and become familiar with the process before progressing to PROD.
At a high level, the plan for moving a server will be as follows:
- Jadu will prepare a new Rocky 9 server instance
- A content and end user registration freeze will be required (we recommend you putting up an announcement on the site for this)
- If you have Jadu Central Forms (XFP), forms and payments will need to be taken offline during this content freeze
- Jadu will synchronise data from the CentOS 7 server to the Rocky 9 server
- An infrastructure switch will redirect traffic to the new Rocky9 server
- The content freeze can be lifted and the CentOS 7 server decommissioned
Will there be any downtime during the move?
Yes, when traffic is redirected in the latter stages of the switchover between servers, we expect a small (1-2 minute) period of downtime where Jadu are in control of your DNS already.
As noted in the above steps, depending on which modules you are actively using, there will be a period of content / end user registration freeze as well as online forms & payments being offline whilst migrating data over to the Rocky server before the servers are switched over.
We host with Jadu - do we need to do anything?
We will contact you over the coming weeks via Jadu Support with a plan for your specific infrastructure and what steps your team will be required to undertake, but largely this will be around:
- Putting up advance notices for your end users and alerting staff of the plan
- Testing / familiarisation of the process with the UAT environment
We host ourselves - do we need to do anything?
If you self-host on Windows Server, you are unaffected.
If you self-host on Linux, we’d recommend moving to a supported operating system. Jadu Central will be supported on RHEL 9 and Rocky 9, see more details on versions below. On both operating systems, we’ll still utilise MariaDB 10.5 (at the time of writing February 2024).
We’ll provide further guidance on the steps required to migrate data in due course should your internal team of systems administrators wish to undertake this process themselves. If you would prefer Jadu to manage this process under professional services, then please contact our Support team.
We’re currently running Jadu CMS x.x.x, do we need to upgrade first?
This depends on the version of the Jadu software your server is currently running.
Rocky 9 supports a minimum MariaDB database version of 10.5. This was first supported in Jadu CMS 21 and XFP 8. Therefore, if your server is running CMS 21, XFP 8 or any version of Jadu Central, we will be able to move you to Rocky 9 as is.
If your server is running CMS 20, XFP 7 or anything prior to this, we will need to upgrade your software version before moving you to Rocky 9. This will be in line with the usual process for software upgrades and managed through the Jadu Support team.
We’re a Hyland Content Portal customer - what happens for us?
Content Portal (CP) is a module on top of Jadu Central (formerly CMS & XFP). As such, any implementations hosted on CentOS 7 will need to be migrated. If you self-host, see ‘We host ourselves - do we need to do anything?’.
The corresponding minimum CP version that can move to Rocky 9 is 6.0.0. If your server is running a version prior to this, you’ll need to upgrade first. See ‘We’re currently running Jadu CMS x.x.x, do we need to upgrade first?’ for more details.
We have another Non-Live site hosted with you - will this move too?
Yes, we’ll communicate separately about moving non-production environments such as DEV/TRAINING etc outside of the traditional UAT/PROD set in due course.
We have a Disaster Recovery service - what will happen with this?
Previously, our DR solution involved having a second site that was always running, and we synced data to this nightly. This meant that it was possible to switch over to a version of the site that was no less than 24 hours out of date at any time.
A limitation of this solution was that, for customers using Jadu Central Forms (XFP), it was necessary for forms to be disabled whilst the site was running from the DR server. This is because the DR site maintains its own copy of the database, so form submissions made between the last sync and a disaster occurring would not be present on the DR site when failed over. Similarly, form submissions during the disaster would not be present when the was once again served from the Live server.
As part of the Rocky 9 migration, customers on the above DR solution will be moved to our improved DR solution. This uses AWS technology to sync data in real time to a disk in a secondary AWS region, which can then be used to launch an instance of the site should a regional disaster occur. Because of the real time sync, the DR site would be much more up to date - never more than a few minutes behind, as opposed to up to 24 hours. This means that it is possible to keep forms online whilst running from the DR site.
For customers already using the real time DR, we will simply ensure that the replication is switched to the Rocky server once this is in place.
We have a PLATINUM resilient infrastructure setup - what will the move involve for us?
Resilient infrastructure behaves slightly differently - it utilises a central database that is accessed by a number of worker servers. We will communicate any tangible differences specific to your infrastructure when we reach out to you, but this will likely simply lessen the content freeze period required compared to GOLD infrastructure setups.
We have a VPN or special networking, will this be affected?
Yes - the VPN will point to your CentOS 7 server. This will either need re-pointing to the replacement Rocky 9 server or a new VPN created for you to use.
We will communicate specifics relating to your VPN when we reach your server in the migration.
We use Jadu Connect - do we need to do anything special for this?
No, Jadu Connect is Multi-tenant SaaS and any maintenance required will be handled by Jadu as part of regular scheduled maintenance as communicated through https://cxm.statuspage.io/.