Hey hbase community, context:
- Streaming millions of entries via apache flink 1.12.2 and hbase connector with integrated phoenix client 5.0.0.7.0.3.0-79 from kafka into hbase - Hbase cluster with v2.0.0.3.0.2.0-50 and 20 worker nodes and 349.15 Regions per Regionserver - database.signal.batch-size=1000 - database.signal.batch-interval=1000 in our system we have a huge amount of logs with "waiting actions to finish on table...." while writing in hbase. These leads to slow processing in flink. Without going much into config details of hbase what are the main reasons that the phoenix client is often waiting until batch writes in hbase are finished? How to tune that? Mit freundlichen Grüßen/Kind regards Frank Scholter Software Developer If you are not the addressee, please inform us immediately that you have received this e-mail by mistake, and delete it. We thank you for your support.
