You can always perform a dry run before the actual deployment to collect
the metrics before going into meaningful deployment.
Other options would be to do a rough math based on the input data set /
complexity of the algorithms used , but that would be an useful
after-step after getting through step 1 to fix the bottlenecks. Until
the actual run / execution of code path of your interest, it would be
hard to arrive at any meaningful conclusion one way or other.
On 01/28/2010 08:58 PM, T. Madana Gopal wrote:
Thank you sir,
'top' helps to find the cpu time related information in run
time.But we need to classify the job as CPU or I/O intensive
before executing it.
Regards,
T.Madanagopal.
top
iostat
utilitities should get some metrics corresponding to the cpu and the io,
that can help identify the nature of the job.
On 1/28/10 3:46 AM, T. Madana Gopal wrote:
Hi all,
What are the factors on which we can classify the jobs given to
hadoop as CPU intensive or I/O intensive?
Regards,
T.Madanagopal.
-----------------------------------------
This email was sent using TCEMail Service.
Thiagarajar College of Engineering
Madurai-625 015, India
Even a stone gets rounded by constant rubbing.
-----------------------------------------
This email was sent using TCEMail Service.
Thiagarajar College of Engineering
Madurai-625 015, India
