SDB:KIWI-LTSP diskless node
tagline: From openSUSE
|This article is being considered for deletion!|
Nothing to see here, obsolete information.
Please do not blank, merge, or move this article, or remove this notice. Refer to this article's discussion page and our deletion policy for more information.
This is the placeholder for Diskless Nodes
This page covers howto setup KIWI-LTSP to work with diskless nodes. You're mainly using the KIWI portion of the script. If you're unaware of what a diskless node is, and the differences between it, LTSP (Thin-Client), and fatclients, check out this wikipedia article: http://en.wikipedia.org/wiki/Diskless_node
Why use Diskless Nodes? Jakob Perry has a blog post going in depth about some pitfalls of using LTSP. He was using LTSP for ~35 clients for day-by-day office work. LTSP works great in many environments, but if you've been seeing performance issues, this page might be for you.
For you, these are some reasons to use diskless nodes:
- Your clients are relatively powerful
- You would like the benefits of centralized management that LTSP offers
- Your systems follow one or a few templates
- You have a larger multimedia need
- You need local access to devices, more than what LTSP provides
Setting up diskless nodes: First, read up on ATA over Ethernet. This demo uses AoE images for deployment. I would suggest setting up a openSUSE / SLED system on an equivalent system first. Put in a hard drive, install the system and configure it to your liking. This will allow you to test it against the hardware your clients will be using. Make sure when installing to NOT have a swap file, and to format ONE partition only. For OpenSUSE 11.1, you should have at least 2GB of ram per client. Once your image environment is setup and tuned to your needs, image installation is fairly straightforward. Take that hard disk and put it within your KIWI-LTSP server.
Setting up the KIWI-LTSP (Deployment) server. will finish tomorrow.