I'll try to reproduce this error as unit test. Meantime I found that number of pages always about 30 (expected 140), on page number 11-13 Scan returns page x10 bigger than page size I asked to return. In this "big" page Scan skip over a lot of rows. It looks like there is some limitation on number of paged rows.
I tried to increase heap memory on client side, run my application as hadoop - jar process, decouple PageFilter and RowFilter, open HBase connection on each page scan, run the same scenario on another cluster. -- View this message in context: http://apache-hbase.679495.n3.nabble.com/Pagination-using-Java-API-wrong-behaviour-tp4068842p4068963.html Sent from the HBase User mailing list archive at Nabble.com.
