Hello, I tried to import a table from hbase 0.90.3 to hbase 0.90.4 on a different cluster by copying the data between those two clusters. I uploaded the data into HDFS and called add_table.rb on that, that finished ok. After that i went to hbase shell and did enable "new_table" which also finished without firing any exception and in the master log I can see that it registered the the request for table enabling and finished ok, saying that all (328) regions of that table are online. Using the UI I can see the new regions being served, but I cannot see the table in the list of available tables in hbase (neither gui nor shell). So when I try to scan that table it obviously says that it does not know such table. When I do disable/enable the new table there is no exception in the master log what so ever, nor regions in transion going wild. I just cannot make it appear in the list of tables.
I have tried hbck and it reports some holes in the new_table's regions and some inconsistencies, I tried check_meta.rb -fix and restarted the cluster but no luck there either. I think that even though there are holes I am running cloudera distro (cdh3u2) installed via packages on debian system. I had some problems with the lzo compression in HBase on the new cluster, but after recompiling and redistributing lzo-hadoop native support those problems went away (it passes the lzo test without firing exception). I also did hcbk on the 0.90.3 cluster and there were no holes indicated for the table. Thanks for your help. -- Stan
