0

Hi in my workplace we are using IBM BigInsights,we create secondary indexes on hbase table using BIGSQL, if the secondary indexes are present the row ingestion speed drops from 3000 rows /sec to 200 rows/sec. How is the performance of phoenix in terms of ingestion speed when indexing is used?what kind of indexing is preferred in phoenix?

siva krishna
  • 23
  • 5
  • how big is the index? is it ordered correctly? have you considered temp table to real table? increased the indexing threads? – Paul Bastide Mar 30 '17 at 12:22
  • Hi,i haven't found anything regarding increasing the indexing threads in IBM BIGSQL. – siva krishna Mar 30 '17 at 13:04
  • its a phoenix setting http://phoenix.apache.org/secondary_indexing.html#tuning scroll to tuning section – Paul Bastide Mar 30 '17 at 14:08
  • i havent tried phoenix,i was saying about the bigsql performance all the time? i was just asking opinion about phoenix, will it be better than BIGSQL? – siva krishna Mar 30 '17 at 14:23
  • an old post, but just curious about what SQL engine you went with (phoenix vs BIGSQL). with later release of BIGSQL version not sure how it going with but what type of ingestion making you a drop in write speed - "LOAD HADOOP" ?, "Load hadoop with bulk load" ? , "Insert..select" ? or "spark jdbc write" ? – Karthick Sep 16 '18 at 20:17

0 Answers0