During last maintenance of Nexus IQ server (update to the latest version and fix some configuration issues) I saw a lot of failes of DevOps pipelines. The root cause was unable to download dependencies from proxy repository that was guarded by Nexus IQ (beacuse I shut down it for maintenance). Later I looked for high-availability configuration of NXIQ but found articles about NXRM HA-C.
Basic load balancing (round robin) is not suitable without database syncing - but there is no way to set up H2 sync on the fly.
So, can anybody share experience or best practice how to organize maintenenace? And how to implement HA or load balancing of Nexus IQ?
HA is not currently available for Nexus IQ Server but it is something that we are actively working on.
Would you be interested to have a discussion around your needs?
Hi @sriram.yeluri,
At the moment there are no expected timelines for this initiative. I would advice contacting your CSE and inform them of your interest. This helps us gauge customers needs and therefore asses our roadmap.
Some clairity: we are working on the technical design and actually coding. Due the to variable nature of software development we don’t have a comitted timeframe. When firm dates are available we’ll be updating folks.
Moving to PostgreSQL is part of this technical design?
I think that DB based on Postgre can help to build HA but I don’t see evidences that stable solution from function and high-load stand point.
The architecture is a standard 3 tier web app like model, where the IQ Server will be as stateless as possible so instances can be independent, making deployment and managment easy.
We are using PostgreSQL to help enable some of the HA capabilities. We already have support for Postgres using a single node IQ Server, which has been successfully adopted by some of our large scale customers.
Much of the tech work now is to ensure multiple IQ Servers can be run independently while still sharing information.
We are already migrating to Postgres SQL now. HA is very much necessary feature as we believe there should be zero downtime when we are upgrading to higher versions.
Same is followed for all other applications in our ecosystem so expecting same for NexueIQ soon.
How I can formally register my vote for this feature?