[
https://issues.apache.org/jira/browse/SAMZA-543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14368338#comment-14368338
]
Chris Riccomini commented on SAMZA-543:
---------------------------------------
Could be. I think something is really weird here, though. Even at 1KB messages,
seeing 17MB/s BEFORE compression seems insanely slow. Especially considering
[~navina] was running this on a laptop with SSDs. We definitely need more
investigation on a different ticket. We are also working on beefing up the
RocksDB JNI test suite (outside Samza), and pushing that upstream. We'll keep
you posted on those results as well.
> Disable WAL in RocksDB KV store
> -------------------------------
>
> Key: SAMZA-543
> URL: https://issues.apache.org/jira/browse/SAMZA-543
> Project: Samza
> Issue Type: Bug
> Components: kv
> Affects Versions: 0.9.0
> Reporter: Chris Riccomini
> Assignee: Navina Ramesh
> Fix For: 0.9.0
>
> Attachments: DisableWAL.diff, SAMZA-543-0.patch, SAMZA-543-2.patch
>
>
> RocksDB uses a write-ahead log by default. This is unnecessary in Samza,
> since we have full durability from a state store's changelog topic. We should
> [disable the
> WAL|https://github.com/facebook/rocksdb/wiki/Basic-Operations#asynchronous-writes]
> in the RocksDB KV store.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
