The above stack trace is related to the same issue that this person is having. The merger task in mapred is trying to load too much into memory at one time. Anyone know if there is any mapred property that controls the size of bytes that the merger tries to do at one time? I suspect this would not be a problem on a large cluster, since tasks are spread out more. But I am running on a single machine with 4gb of memory (moved to a linux machine with more memory). I can not get more memory then this!
http://mail-archives.apache.org/mod_mbox/hadoop-mapreduce-user/201006.mbox/%3c312639.9108...@web114409.mail.gq1.yahoo.com%3e -- View this message in context: http://lucene.472066.n3.nabble.com/Tika-Excel-parsing-causing-out-of-memory-tp1188201p1224730.html Sent from the Nutch - User mailing list archive at Nabble.com.
