There are multiple possibilities for this. 1. Apache Drill currently is supported for Hadoop 2.7.4 and not yet for Hadoop 3.x. 2. The Parquet file is of a later version. Drill supports Parquet 1.10.0 3. Your Apache Drill version is quoted as being 1.5 . I'm guessing that is a typo and you meant 1.15
>From the error message (since it is a SYSTEM error), it is most likely #1. I'm >not sure erasure coding can play a role, since the FS should be abstracting >the erasure coding and the APIs for reading blocks from the file would >(probably) be unchanged between the Hadoop 2.x and 3.x. ~ Kunal On 4/15/2019 11:45:22 AM, jinhwan choi <[email protected]> wrote: Hi Drill Team I upgrade my cluster HADOOP 3.1。 (apache drill is ver1.5) And, I enabled Hadoop 3.X erasure encode feature on this cluster. (and made some new parquet file) But after upgrade, I faced this error. -------------------------------------------------------------- SYSTEM ERROR: BlockMissingException: Could not obtain block: BP-124827589-172.23.55.207-1483001524978:blk_-9223372036854744080_1954863 file=/drill/data/log_568803_R_521_2019040510190000.parquet Please, refer to logs for more information. [Error Id: 940a3b08-0629-400e-8cfb-fd297e1730c7 on tst-nss-hdn06:31010] -------------------------------------------------------------- Can anyone please let us know the solution for this or if there are any alternative. Regards, Choi
