Whiteboard: Scalable NFS File Serving with Cluster File System
davidnoy
July 6th, 2009
Storage Foundation Cluster File System (CFS) delivers modular capacity scalability and performance with improved reliability. David Noy, Sr. Product Manager for CFS, uses a whiteboard “chalk talk” to describe the optimal implementation architecture for CFS and how to get the most from it. He also describes the differences between CFS and Storage Foundation Scalable File Server (SFS) and why IT organizations might want to use one or the other.
Video Upload:
How does client distribution happen?
Hi David,
Thanks for the presentation. I have taught the CFS course a couple of times to customers as a Symantec instructor, so am familiar with most of what you are referring to, and I think you've done a great job summarizing how CFS works in an NFS environment. I think the white paper here (http://eval.symantec.com/mktginfo/enterprise/white_papers/ent-whitepaper_cluster_file_performance_scalability_12-2007_13579374.en-us.pdf) does a pretty good job of backing up the statistics you are referring to in regards to scalability.
One question I have: When you say 'clients connect randomly' to those servers, how is this accomplished? Obviously each server needs an independent virtual IP address, so is this accomplished by having a DNS round-robin function (outside of VCS control) in the front? Or is there something else about NFS that allows this distribution of clients to happen?
Cheers,
Tomer
Tomer, Thanks for teaching
Tomer,
Thanks for teaching the CFS classes. Yes, the way it works is to have multiple virtual IPs in DNS and have the DNS round-robin help redicrect cleints to different nodes randomly.
thanks,
- KR.
Would you like to reply?
Login or Register to post your comment.