Oracle Database 23.4 Free now available on gvenzl/oracle-free

Last Thursday (2nd of May 2024), Oracle released Oracle Database 23ai on Oracle Cloud but included Oracle Database 23ai Free for on-premises. Some of you may have already heard that 23ai is the new name for 23c, so what you get is still Oracle Database 23.4 and it’s the same code base as 23.3 and 23.2 before. The difference is that 23.4 comes with a lot more AI-related features, given last year’s change in focus in the industry, such as a new VECTOR data type and Vector Search. Anyway, that’s probably worth a blog post on its own. But it is important for the users of gvenzl/oracle-free to understand that 23ai is not a brand new release that may need some migration or code changes in your app, it’s simply the new name for what was supposed to be 23c.

So, thanks to a rainy Saturday yesterday, I took the time and refreshed https://hub.docker.com/r/gvenzl and https://github.com/gvenzl/oci-oracle-free/pkgs/container/oracle-free/ to now also provide Oracle Database 23.4 Free. As usual, the slim, latest, full and 23-slim, 23-latest, 23-full have all moved forward to 23.4. So, for those of you who have already used one of these tags, congrats, your integration tests are already running on 23.4. For those of you who use a 23.3 or 23.2 specific tag, I highly recommend moving to 23.4 or one of the RU number agnostic tags (slim, 23-slim, etc.)

To make the older versions a bit more apparent, you will now see the following output in your container log if you are running on an older version:

################################################
NOTICE: YOU ARE USING AN OLD IMAGE VERSION 23.3!
PLEASE CONSIDER UPGRADING TO THE LATEST VERSION!
################################################

Another change you may notice is that the faststart images no longer contain the layers from the non-faststart images. That’s due to a strange issue I found when upgrading buildah that I can not yet explain, see https://github.com/containers/buildah/issues/5491. Hence, for now, I have to rebuild the non-faststart images once the faststart ones “swallow” the original non-faststart images. (If anybody has any ideas or pointers, I’m all ears.) However, if you don’t care about having both images on your machine and the time it takes to download them, then you really don’t have to worry about this.

So, the easiest to get a brand new Oracle Database 23ai Free image via docker / podman:

podman run -d --name 23ai -p 1521:1521 -e ORACLE_RANDOM_PASSWORD=y -e APP_USER=vector -e APP_USER_PASSWORD=vector gvenzl/oracle-free:slim

Have fun, and, as always, thanks for using these images! It does make the time spent on rainy Saturdays way worthwhile! 🙂

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.