siying commented on code in PR #40981:
URL: https://github.com/apache/spark/pull/40981#discussion_r1180870219
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -247,14 +253,7 @@ class RocksDB(
}
def prefixScan(prefix: Array[Byte]): Iterator[ByteArrayPair] = {
- val threadId = Thread.currentThread().getId
- val iter = prefixScanReuseIter.computeIfAbsent(threadId, tid => {
Review Comment:
I would follow your intuition that removing the optimization won't cause a
big issue. I guess if we would keep it, the logic might be more complicated, so
it makes sense, so it makes sense.
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/RocksDB.scala:
##########
@@ -61,17 +60,27 @@ class RocksDB(
// Java wrapper objects linking to native RocksDB objects
private val readOptions = new ReadOptions() // used for gets
- private val writeOptions = new WriteOptions().setSync(true) // wait for
batched write to complete
Review Comment:
It doesn't matter but for a record, I don't see a reason why we need to do
sync in the first place. WriteOptions.sync is only there to get data preserved
after machine power reset.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
