Reviving the cluster question
-
The last posts I saw regarding this were about a year ago, and I have a need for this so reviving the topic. I saw some mention that HA is on the roadmap - is there a publicly available roadmap I haven't found, or are there any updates on that effort?
Currently our primary need is for fail-over, although as needs grow over the coming few years we will likely need to implement an HA solution at some point.
If there's no official guidance on this yet, is there anyone out there that can share a successful implementation of either an HA or fail-over implementation?
Ultimately we will have some individual apps that will incorporate their own HA/fail-over solutions, but with our current size and needs implementing that at the Cloudron host level for now would be a hugely easier path.
-
There is no HA in Cloudron . Can you tell us a bit about what HA looks like to you? Are you thinking of handling situations where the server will just stop working? AFAIK, big VPS providers already have this built in to their systems.
I have to say I have not had VMs fail for me in a decade now. The only memory I have is SSDNodes VM constantly becoming unresponsive , but they are not really meant for production stuff since they are giving stuff for such low prices.
-
Cloudron is generally great at what it does, however we've had a few instances of the server becoming unavailable for one reason or another, and while it's unresponsive and/or waiting for attention, it's associated apps likewise become unavailable. If HA in general is not an option, at least having some fail-over means would be a great secondary option.
-
To me it would be a selfhosted solution / load balancing peace of mind thing. That said I haven't really seen the server struggle to do anything for my less than 10 connections simultaneously. I'm not sure what the capacity for a single server would be, if I avoid apps that are CPU heavy, just use nextcloud for file sharing.
