Am looking to manage more than 8000 hosts through puppet config mgmt, is foreman can mange this, if not please refer a right tool
Obviously it depends on how you will use the tool, how big your resource catalogs are etc. But we’ve seen much bigger setups successfully using Foreman, so the short answer is - yes.
Obviously it can, at eBay/Paypal it was reported to manage 100,000 nodes. However this does not come free, be prepared to invest a lot of resources into this - you need skilled ops, network engineers, good managers, good planning, software and hardware.
I suggest to start small and grow big. Do a decent POC. And if you are unsure you can always hire consultant(s) or buy Foreman-based products.
We at Dell are managing ~ 5000-6000 nodes. We dont run every 30 mins however, so that may change things. But we have ~ 200 custom facts and all nodes check in every 6 hours. No issues here.
A blog post! A blog post! A blog poooosst please.
There had to be some quicks here and there you needed to solve. We want more feeedback really.
Thank you everyone for the quick response. We are planing to running a base configuration for windows on all nodes and run once in a day is good enough. Based on @lzap @Jeff_Sparrow inputs, I can try managing through foreman itself. I will start building the base and will come back with future questions. once again thanks you everyone.
We do 16k nodes across 15 environments, approx 800 classes per environment. Check-ins every 30 minutes for puppet and “as needed” with ansible.
Comes at a cost though, 18 VM puppet masters @64gb RAM and 16 cores each, which run pretty high usage overall (30 to 50 percent cpu with spikes higher as I take multiple puppet masters out of load balancer)
Definitely doable. 100k at eBay is insane, I’ve seen the presentation, but wish there was more specifics about “how” to scale to that level… Tweaking things and perfoormance tuning becomes tough!
Thanks, I created a new table with your info, would you mind updating it a bit? Tutorial: Hardware requirements for Foreman deployments