hadoop - GC overhead limit exceeded container killed in Pig -

i executing 13 table map join on our development environment in hadoop 2 cluster using yarn.all table join left outer main table .total number of there 15 join join.

since of small table less 200-300 mbs,so used using 'replicated' execute script code.it executes quite fast stuck past 95-99% .when check application url ,2 reducers failed throwing error " gc overhead limit exceeded container killed applicationmaster.container killed on request.exit code 143 container exited non-zero exit code 143".

other reducer failed error timed out after 300 secs container killed applicationmaster.container killed on request.exit code 143..

you can tweak values directly within pig.

in application url, check job properties, , see current values of

mapreduce.map.memory.mb    mapreduce.reduce.memory.mb mapreduce.map.java.opts mapreduce.reduce.java.opts mapreduce.task.io.sort.mb 

you can begin tweaking raising values of properties 512 @ time. however, not raise value of mapreduce.task.io.sort.mb on 756.

to change values directly pig, @ top of pig, before code, add properties follows :

set mapreduce.map.memory.mb    3072; set mapreduce.reduce.memory.mb 3072; set mapreduce.map.java.opts -xmx2764m; set mapreduce.reduce.java.opts -xmx2764m; set mapreduce.task.io.sort.mb 756; 

the above properties worked me data-set of similar size yours.