DSpace Migrations to Amazon Linux 2 (Spring 2022)

As part of routine maintenance of its hosted repositories, TDL will migrate all hosted DSpace repositories to a new operating system within the Amazon Cloud, from Amazon Linux 1 (AL1) to Amazon Linux 2 (AL2). To do this, TDL will migrate your repository and its contents to a new server with the AL2 operating system, along with updated underlying software systems like PostgreSQL and Tomcat.

Will my repository be down during the migration?

No. Your site should experience no downtime during this process. The site will remain publicly available and users will be able to search, browse, and download items. However, TDL will make your current DSpace repository READ ONLY at the start of your maintenance window, so that no changes (e.g. new submissions) are possible. This will prevent any new data from failing to copy over to the new server.

What will users notice after the migration?

Not much, if anything. We are not upgrading the DSpace software itself, and no new DSpace features will be included. You may notice improvements in performance (faster loading time, for instance) because of the improved operating system, database software and web server software.

If you are interested in the details, here’s what is changing with the software systems underlying and working with DSpace:

  • Upgrading PostgreSQL from version 9.6 to 13

  • Upgrading Tomcat from version 8.5 to 9.0

  • Upgrading Shibboleth from version 2.6 to 3.3

  • Moving from Amazon Linux 1 to Amazon Linux 2 operating system

What will happen during the migration?

TDL will make your current DSpace repository READ ONLY at the start of your maintenance window, so that no changes (e.g. new submissions) are possible. This will prevent any new data from failing to copy over to the new server.

The start of your maintenance window will be on or around 3pm* until the following morning. During that time, your repository will be up and available, and users can search and download items, but no one will be able to submit or make changes to items.

The migration itself will take several hours. The following morning TDL staff will make sure the migration completed successfully and make the final changes (point your URL to your fresh new repository, for example), and remove the READ ONLY restriction. You will hear from TDL’s DSpace tech lead Nick Woodward when the process has finished.

*For most TDL-hosted repositories, we will start the maintenance window at 3pm. But please note that if you have a large repository (over 400 gb), the data migration will take longer, and we may need to start the migration earlier in the day (at 1pm for instance). We will let you know on the day of your migration when we plan to start the process.

When will my repository be migrated?

We are aiming to migrate 3 repositories per week, starting Monday, March 7.

The order of the migrations can be seen here:https://texasdigitallibrary.atlassian.net/wiki/spaces/DIGREP/pages/2990899220 . Because of uncertainties around disruptions that may cause delays, we are not specifying a date for each repository migration, but you can estimate the time frame for your migration based on this tentative information.

TDL will inform you via email on the morning of the day we perform the migration. If you have unavoidable conflicts that make it disruptive for the maintenance to happen on that day, please let us know asap.