[
https://issues.apache.org/jira/browse/KAFKA-9450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17018781#comment-17018781
]
Guozhang Wang commented on KAFKA-9450:
--------------------------------------
I think John's idea is to e.g. use a preserved key to store the current offset
of the changelog as a preserved key, so that each time we update a value, we
will also update this key for offset update. By doing this we would not need a
rocksDB#flush when checkpointing, and neither do we need a checkpoint file. One
tricky thing however, is that the offset of the written changelog is only
updated and known after the batching produce response returns, which may not be
immediate after the RocksDB value update.
> Decouple inner state flushing from committing with EOS
> ------------------------------------------------------
>
> Key: KAFKA-9450
> URL: https://issues.apache.org/jira/browse/KAFKA-9450
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Sophie Blee-Goldman
> Priority: Major
>
> When EOS is turned on, the commit interval is set quite low (100ms) and all
> the store layers are flushed during a commit. This is necessary for
> forwarding records in the cache to the changelog, but unfortunately also
> forces rocksdb to flush the current memtable before it's full. The result is
> a large number of small writes to disk, losing the benefits of batching, and
> a large number of very small L0 files that are likely to slow compaction.
> Since we have to delete the stores to recreate from scratch anyways during an
> unclean shutdown with EOS, we may as well skip flushing the innermost
> StateStore during a commit and only do so during a graceful shutdown, before
> a rebalance, etc. This is currently blocked on a refactoring of the state
> store layers to allow decoupling the flush of the caching layer from the
> actual state store.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
