Hi Andreas,
From memory, some of the issues we had in Docker were around permissions, and some of the default restrictions. I have very limited Docker exposure, although, others working with me were much more experienced with it. So, I was under the arguably misguided impression that an on-prem environment would easily fit into a standard Docker container.
We had better success with VirtualBox, although, VirtualBox is not a container. However, LXC/LXD (which is a container), appeared to work out-of-the-box with a lot less hassle than Docker.
From my limited understanding, LXC/LXD does seem a better fit for containerizing Informix due to it's wider scope (as opposed to Docker's more app-focused approach). There could well be an Enterprise capable Docker image for Informix that's available for people to use (if you know of one, or have setup such an image yourself, I'm sure it would be greatly appreciated).
Regards,
Daniil
Original Message:
Sent: Mon July 13, 2020 05:30 AM
From: Andreas Legner
Subject: Issue: shared memory not initialized for INFORMIXSERVER 'informix'
Hi Daniil,
shouldn't this only be a matter of bringing the 'edition installer' for the desired (and acquired) edition onto the Docker instance and applying it there?
I'd not see how Informix in the docker image should be different from (the very same) Informix installed 'on prem'.
BR,
Andreas
------------------------------
Andreas Legner
------------------------------
Original Message:
Sent: Mon July 13, 2020 02:00 AM
From: Daniil T
Subject: Issue: shared memory not initialized for INFORMIXSERVER 'informix'
Good luck with the Docker adventure.
We had a number of issues trying to get Docker to work with our particular database. In the end we just went with LXD, which so far seems to be a lot more accommodating.
We did get docker to work in the end, using a Developer Edition license, and a dataset that complied with the license storage restrictions. But, the experience put us off using docker for any significantly sized datasets.
If someone somewhere has been able to create an image that can successfully work with a production sized dataset, I'd be interested to know the details of how it works.