Github user tdas commented on the issue:
https://github.com/apache/spark/pull/20958
We have made it clear that sockets is ONLY for testing and will not recover
data from checkpoints. So I see no problem that it throws errors when
attempting to recover. May we can improve the error message by making it clear
that recovery is not supported.
If you indeed want to forget lost data and proceed, then that should be an
opt-in. We could do this by explicitly setting a source option (like
failOnDataLoss = false in Kafka source).
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
