hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "tanjun_2525"<tanjun_2...@163.com>
Subject 回复: RE: How can I know how many mappers created?
Date Mon, 08 Aug 2011 05:25:22 GMT
Hi Subroto,
Thank you very much!

2011-08-08



tanjun_2525



发件人: Subroto Sanyal <subrotosanyal@huawei.com>
发送时间: 2011-08-08 12:33
主 题: RE: How can I know how many mappers created?
收件人: mapreduce-user@hadoop.apache.org



Hi Jun,
 
Request you to go through this link:
http://wiki.apache.org/hadoop/HowManyMapsAndReduces
 
Following configuration may help you :
mapred.min.split.size
mapred.map.tasks (This is just a hint to InputFormat)
 
But how can I know how many mappers have been created while doing my job?
The implementation of InputFormat actually defines how many Mappers are going to be in the
submitted job.
 
Can I do some sets to make it only 2 mappers created and each mapper runs on one node?
The scheduler implementation has the logic to schedule mappers and reducers.
 
 
 
Regards, 
Subroto Sanyal



From: 谭军 [mailto:tanjun_2525@163.com] 
Sent: Sunday, August 07, 2011 12:15 PM
To: mapreduce-user@hadoop.apache.org
Subject: How can I know how many mappers created?
 
Hi,
I just want 2 mappers created to do my job for there are only 2 data nodes.
I think that is the most efficient.
But how can I know how many mappers have been created while doing my job?
Can I do some sets to make it only 2 mappers created and each mapper runs on one node?
Thanks


--
Best regards!
Jun Tan
Mime
View raw message