[
https://issues.apache.org/jira/browse/RATIS-872?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tsz-wo Sze reassigned RATIS-872:
--------------------------------
Assignee: Tsz-wo Sze (was: Lokesh Jain)
Issue Type: Improvement (was: Bug)
[~ljain], since there are no activity for a long time, you probably are not
working on this. Lets me do it.
> Ratis retry cache improvements
> ------------------------------
>
> Key: RATIS-872
> URL: https://issues.apache.org/jira/browse/RATIS-872
> Project: Ratis
> Issue Type: Improvement
> Components: client, server
> Reporter: Lokesh Jain
> Assignee: Tsz-wo Sze
> Priority: Major
> Attachments: Ratis Retry Cache Improvements.pdf
>
>
> The duration for which a client request entry is kept inside the retry cache
> should be long enough to cover all the retries made by the client. Retry
> cache would need to keep the metadata for all client requests until that
> duration is over. It can therefore require a large amount of heap in order to
> handle client requests. Since active heap usage is high it can also increase
> the frequency of JVM garbage collection.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
