Kylin only manage the life cycle of the job. However the job runs and fail over depends on the hadoop cluster behind.
We haven¹t met such case. Probably it is due to inconsistency in hbase(where we store all the metadata like jobs, tables). You may want to check the htable kylin_metadata(see kylin.metadata.url in kylin.properties) Best Regard Zhou QianHao On 5/18/15, 1:39 PM, "Tao Wong" <[email protected]> wrote: >how do kylin do when this scenario occur ? >as we know , the machine may crash while the cube is building. > >has anybody met this ? > >2015-05-18 12:21 GMT+08:00 Tao Wong <[email protected]>: > >> Hi all >> >> I closed kylin by mistake when there are still jobs were running. >> I restarted it and found nothing ,no cubes,no jobs,no tables. >>
