[jira] [Commented] (CASSANDRA-14309) Make hint window persistent across restarts

["Stefan Miklosovic (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-14309?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17432882#comment-17432882 ] 

Stefan Miklosovic commented on CASSANDRA-14309:
-----------------------------------------------

I would like to move this forward regardless of the docs status, I can deal with that later / create another ticket to track that. I have rebased this stuff on top of the current trunk after big patches for Benedict, all that took was basically to reuse global clocks functionality in places we are using System.currentMillis(). 

I have also addressed the comment from Mick about making sure the map with earliest hints is not growing constantly by removing it on buffer's flushing to disk.

build is here: https://ci-cassandra.apache.org/job/Cassandra-devbranch/1225/#showFailuresLink 
failures seem to be unrelated to this and other recent builds are failing on them too.
circle ci:  https://app.circleci.com/pipelines/github/instaclustr/cassandra/591/workflows/c696306b-60f6-4308-9387-e5ade8fb9470
dtest branch: https://github.com/apache/cassandra-dtest/pull/153

quick check from [~brandon.williams] and [~mck] and at least one +1 would be appreciated.

> Make hint window persistent across restarts
> -------------------------------------------
>
>                 Key: CASSANDRA-14309
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-14309
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Consistency/Hints
>            Reporter: Kurt Greaves
>            Assignee: Stefan Miklosovic
>            Priority: Low
>             Fix For: 4.1
>
>          Time Spent: 1h 50m
>  Remaining Estimate: 0h
>
> The current hint system stores a window of hints as defined by {{max_hint_window_in_ms}}, however this window is not persistent across restarts.
> Examples (cluster with RF=3 and 3 nodes, A, B, and C):
>  # A goes down
>  # X ms of hints are stored for A on B and C
>  # A is restarted
>  # A goes down again without hints replaying from B and C
>  # B and C will store up to another {{max_hint_window_in_ms}} of hints for A
>  
>  # A goes down
>  # X ms of hints are stored for A on B and C
>  # B is restarted
>  # B will store up to another {{max_hint_window_in_ms}} of hints for A
>  
> Note that in both these scenarios they can continue forever. If A or B keeps getting restarted hints will continue to pile up.
>  
> Idea of this ticket is to stop this behaviour from happening and only ever store up to {{max_hint_window_in_ms}} of hints for a particular node.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org