Hi Adam, What version are you running, and are you using multiple reducers in your HFileOutputFormat job? There was a bug in 0.20.3 which caused this case to produce somewhat broken tables.
-Todd On Mon, May 24, 2010 at 10:55 PM, Adam Silberstein <[email protected]>wrote: > Hi, > A colleague and I are working on testing a few HBase features, notably bulk > import (mentioned in > > http://hadoop.apache.org/hbase/docs/current/api/org/apache/hadoop/hbase/mapr > educe/package-summary.html) and running M/R jobs using HBase as input. > > We¹re taking the following steps: > 1a. Load HBase with a M/R job using the normal API. > OR > 1b. Load HBase with bulk import. > > THEN > > 2a. Using the shell, do a ³count² over the table. > OR > 2b. Run a M/R job that scans the whole HBase table (and nothing else). > > Of the 4 combos, 3 are fine: 1a+2a, 1a+2b, 1b+2a. We¹re having trouble > with > 1b+2b. When we run the M/R job, it doesn¹t seem to read in any records, > but > there are no explicit errors in either the Hadoop or HBase logs. > > This seems odd. It shouldn¹t matter how we load the table, and the shell¹s > count operator seems to work correctly either way, counting all the > records. > The M/R job in 2b is the same no matter how we load the table. Any ideas > on > what might be wrong with the bulk import to cause this problem? We¹re > thinking maybe something with the region boundaries, although they look ok > in the GUI. > > Thanks for any suggestions, > Adam > -- Todd Lipcon Software Engineer, Cloudera
