Can you give the whole stack trace for WrongRowIOException ? Was the cluster running Export using the same version of hbase ( 1.0.0-cdh5.5.2) ?
Thanks On Mon, Nov 21, 2016 at 4:35 PM, Julian Jaffe <[email protected]> wrote: > Hbase Version: 1.0.0-cdh5.5.2 > > We're importing the data using `hbase > org.apache.hadoop.hbase.mapreduce.Import 'table.name' /path/to/backup` > (The data was exported from an HBase instance on another cluster using > `hbase org.apache.hadoop.hbase.mapreduce.Export` and then distcp'd between > the clusters). > > On Mon, Nov 21, 2016 at 4:29 PM, Ted Yu <[email protected]> wrote: > > > I did a quick search - there was no relevant JIRA or discussion thread at > > first glance. > > > > Which hbase release are you using ? > > > > How do you import the data ? > > > > More details would be helpful. > > > > Thanks > > > > On Mon, Nov 21, 2016 at 2:48 PM, Julian Jaffe <[email protected]> > > wrote: > > > > > When importing data into a fresh HBase instance, after some time the > > import > > > throws the following exception: > > > > > > Error: org.apache.hadoop.hbase.client.WrongRowIOException: The row in > > > \x00\x00\x0767341283611_10153807927108612\x00\x80\x00\ > > > x00\x00\x84)L\xA7/IN:nme/1461847340445/Put/vlen=42/seqid=0 > > > doesn't match the original one > > > \x00\x00\x0767341283611_10153805927108612\x00\x80\x00\ > x00\x00\x84)L\xA7 > > > > > > (The non-matching row differs on different runs). > > > > > > If the import is allowed to run to completion, the row count of the > data > > > imported is less than the row count of the source data. > > > > > > Googling for this error only turns up the source code that generates > the > > > error, so it doesn't seem to be a common problem. > > > > > > Can anyone provide any guidance? > > > > > > Julian Jaffe > > > > > >
