I'm going to build a cluster lab to test solutions and new applications for use in a larger cluster considered to be highly available, scalable and fault tolerant. I want to know available open source solutions for such a cluster lab, any proposal or comment is appreciated.
-
Product and service recommendations are off topic per the updated FAQ.– sysadmin1138 ♦Aug 9, 2012 at 17:12
Comments disabled on deleted / locked posts / reviews
|
1 Answer
There are quite a few different solutions you could go with, depending on what you ultimately are trying to accomplish ( and I would even try to pick one that mimics the big one)
Most common might be Beowulf Clusters. Other Parrellel ones are PVM or MPICH2. Or you can even custom write some software that offloads task to a set of known machines. (I know Gentoo does this with distcc to distribute the task of compiling software).
I think it really depends on what you are trying to do / mimic.
For data intensive, check out Apache's Hadoop it's an open-source project of utilities, file system and mapreduce implementation
-
My applications will be data-intensive and optimised to have high throughput typically for file I/O. Applications developed using MapReduce or for Content Delivery Networks are such ones.– eagleApr 10, 2011 at 5:45
-
Might want to check out Apache's Hadoop. Adding link in my answer Apr 10, 2011 at 16:15