Mark Minasi's Tech Forum
Sign up Calendar Latest Topics

  Author   Comment  

New Friend (or an Old Friend who Built a New Account)
Posts: 56
Reply with quote  #1 
Or how is it going? We have 4 or 5 Hyper-V clusters and the tech doesn't seem too far off of that (Since it sorta like using C:\ClusterStorage as a share...). Anyway, what do you like? What DON'T you like? What are the gotcha's?

We have a few Virtual File servers that get absolutely HAMMERED and we need something a bit more robust than the current single point of failure.

EDIT: Also, we use DFS. How does the CAFS fit into that? If we have these burly CAFS servers as the front end for a folder in DFS, can the physical hosts act as Namespace servers too for a particular namespace? That way if one went down the other could pick it up?

If Chewbacca lives on Endor - You must acquit!

New Friend (or an Old Friend who Built a New Account)
Posts: 69
Reply with quote  #2 
To directly answer your question: no..

The longer answer is;
- we have a few w12r2 cluster but for cafs to work you need w12r2+ clients and ours are still ws08r2 because we think ws12 rdsh does not scale enough (long story).
- we do however are employing s2d clusters, which could solve your problem of hammered file servers as its designed to scale to millions of iops and uses distributed storage instead of shared storage. You can find more on (technical, no marketing stuff).
- s2d can be build on top of sofs, but we decided on conveged to the hyperv role is on the s2d nodes.

We decided on s2d because of the same problem, hammered file servers as there is a btrieve (novell's 3.x main file database tech) engine running on it so we need large open file handles (400k+) and lots of iops. This design (theorticly) should give us much flexibility depending on the demand (storage vs iops).

Home is where is sleep
Previous Topic | Next Topic

Quick Navigation:

Easily create a Forum Website with Website Toolbox.