[
https://issues.apache.org/jira/browse/RATIS-533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16824615#comment-16824615
]
Josh Elser commented on RATIS-533:
----------------------------------
Going through the current list and recording here for posterity:
* {{RaftServerConfigKeys.Log.setMaxCachedSegmentNum}} defaults to 6, Ozone sets
this to 2. Don't know a reason why we'd need to change this now.
* {{RaftServerConfigKeys.Log.StateMachineData.setSync}} defaults to true, don't
need to call.
* {{RaftServerConfigKeys.Snapshot.setAutoTriggerEnabled}} defaults to false,
Ozone sets to true. Don't understand the implications of it yet.
* {{RaftServerConfigKeys.Log.setElementLimit}} Default of 4096 limit on "queue
entry", not sure what queue this is
* {{RaftServerConfigKeys.Log.setByteLimit}} default of 64MB limit on "queue
size", not sure what queue this is.
* {{RaftServerConfigKeys.Log.setWriteBufferSize}} default to 64KB, looks like
we could bump this up to amortize some write costs?
* {{RaftServerConfigKeys.Log.StateMachineData.setSyncTimeoutRetry}} defaults to
retrying indefinitely. Ok for now?
* {{RaftServerConfigKeys.Log.StateMachineData.setCachingEnabled}} defaults to
false, Ozone sets to true. Don't understand implications of it.
> Set important server config options for logservice
> --------------------------------------------------
>
> Key: RATIS-533
> URL: https://issues.apache.org/jira/browse/RATIS-533
> Project: Ratis
> Issue Type: Improvement
> Components: LogService
> Reporter: Josh Elser
> Assignee: Josh Elser
> Priority: Major
>
> Mukul was kind enough to link to
> [https://github.com/apache/hadoop/blob/d31c86892e0ceec5d642f76fc9123fac4fd80db8/hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/common/transport/server/ratis/XceiverServerRatis.java]
> in Ozone which lists a bunch of RatisServer configuration keys that Ozone
> sets.
> We should go through this list and set the ones that make sense for the
> LogService too. For example, initial testing with a longer leader election
> timeout helps significantly.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
