Replies: 1 comment 2 replies
-
There's an old wiki page about HA installations https://github.com/ckan/ckan/wiki/CKAN-High-Availability. But typically tech team does not have resources to maintain complex documentation, even the current "this is how you can get CKAN running on ubuntu" takes an effort. There are at least 3 major cloud providers and multiple smaller ones which each do things slightly differently. If we have documentation for it, we will end up with support requests which we don't have resources to resolve. And then you have to account the security too, as many of the users are not experts in server maintenance, even less is cloud maintenance. |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
It's quite feasible to split a CKAN installation over multiple servers. For example, we have an auto-scaling group of servers behind a load balancer to run CKAN, with a second auto-scaling group of servers that run the job workers, and Solr installed on a third set of servers (not so easy to auto-scale, but it's at least isolated so that it isn't competing for CPU and memory), and connecting to Redis in the cloud.
Should there be some documentation of what's needed to make this work, at least for a source install (which is intended as the most flexible approach)?
Beta Was this translation helpful? Give feedback.
All reactions