I don't not believe that S3 is currently a supported filesystem for transactional tables. I believe there are plans to make this so.
On 12 June 2018 at 17:50, Abhishek Raj <abhishekra...@yahoo.com> wrote: > Hi. I'm using HiveBolt from Apache Storm to stream > <https://cwiki.apache.org/confluence/display/Hive/Streaming+Data+Ingest#StreamingDataIngest-StreamingRequirements> > data into a transactional hive table. It works great when the table is > backed by hdfs, but starts throwing error when the location is s3. The > error I get is, "No filesystem for scheme: s3" > > Just wondering if it's possible to stream data into a s3 backed hive table > from storm or is hdfs the only filesystem supported. > > Any insights would be great. > > Thanks. >
