How to monitor and compute performance on rocks cluster?
Hello everyone,
After a lot of googling, I've setup a rocks cluster with 1 frontend node and 2 compute nodes. The rocks version is 6.1 and packaged centos version is 6.5. The cluster is built and the nodes are up and running. I am unable to find good "How to's" to run a service on rocks, monitor the cluster, which node is getting the maximum load? Are the processes being distributed across all the nodes and so on. I tested small jobs like compiling a c program using mpicc and running the same using mpirun. Which worked. Does a tool specifically to monitor the cluster exists? a tool, for example, generates unnecessary load, a benchmark, or something. What are the good guides i can follow? Please throw some light. Thanks! |
Quote:
http://central6.rocksclusters.org/ro...on/base/6.1.1/ The Rocks documentation has ALL the commands with explanations.. |
@TBone,
I did go through the entire document. There's no information about performance measurement, parallel processing, proving the fact that the cluster is better than a single computer, which is exactly what i asked. |
Quote:
http://etutorials.org/Linux+systems/...I+Performance/ https://groups.google.com/forum/#!to...rs/ULIeg77KMwk ...like those two, which reference many OTHER tools. Rocks has built-in utilities you can use (referenced in the documentation), which you can script with. There are also third-party tools, like those mentioned above. |
@TB0ne,
Thank you so much for the links! Got what i wanted :) |
Quote:
|
All times are GMT -5. The time now is 08:49 AM. |