CODEWITHSUNDEEP
nutch
sriram
sriram

Reputation: 732

Nutch Fetcher: Number of urls fetched persecond

I would like to know a property which denotes how many urls to be fetched per second. I have seen from net that it is actually number of map tasks * number of threads. But in my case the numbers don't match when I verify them. Hence is there any property to specify the number of urls to be fetched per second?

regards,

V.Sriram

Upvotes: 1

Views: 580

Answers (1)

sriram
sriram

Reputation: 732

I figured out this problem myself. Actual fetch rate is dependent on the number of map tasks that run throughout the fetcher time limit and the thread per host property. Also the number of fetcher threads play a part here.

For example, if the total number of map tasks is 8 and the number of domains in the input urls is 3, then only 3 map process will run in that fetch cycle. And if each map process has enough urls from the same domain that it would run for the fetcher.timelimit.mins then the total fetch rate of the system will be the sum of the fetch rates of these individual map tasks.

Upvotes: 2

Related Questions