Hi,
I’m incredibly frustrated with Rancher. We’ve been trying to use Rancher for 2+ months, so I’d say I"m not a newbie.
But, 100% of the time, our deployments fail. Every deployment of a new software release burns 8+ hours, and we’re a small engineering team, so it’s very materially impacting.
We have 15 or so stacks, each with 1 to 10 services in it. Each of these services are essentially running the same core “app” image–working off different queues.
In all cases, an attempt to use rancher -compose to push an upgrade fails 100% of the time. This failure leaves Rancher server in a totally busted state, and we literally have to go to each host, clean it, go to rancher server blow it away and rebuild it.
Every. Single. Time.
I’ve literally burned 100+ hours ONLY working on rancher related configuration. We’ve written custom scripts to build our rancher-compose files and do do the deploy. But, rancher fails all the time.
This time, by the way, it seems the deploy was stuck. The web front end states it was downloading for 2+ hours a FS layer. And, when we went to the hosts (which it says were up and showed it was generating traffic/cpu/graphics), the docker images were NOT running (though rancher server showed them as running).
I need to know if this (Rancher) is really something real, or a waste of our time. We’re considering reverting to a 10 to 20 line bash script to do deploys… as that would do it more reliably than Rancher at this point.
Chat rooms (IRC) are unresponsive…
I’m at amy wits end, and could use some guidance here. I’m frankly tired of burning ALL my time working on deploying my code using rancher, versus writing and developing code which makes me money.
Anyone?
Again: 1.2.0pre3 Rancher, Dedicated hosts and hardware (64GB RAM 8 core machines with SSDs), GB ethernet local switch connectivity. Bare metal. It doesn’t get any cleaner/tighter.
