[
https://issues.apache.org/jira/browse/CASSANDRA-11168?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joel Knighton updated CASSANDRA-11168:
--------------------------------------
Status: Open (was: Patch Available)
This doesn't look quite right to me yet.
The metric {{incrPastWindow}} pre-patch simply records all the times shouldHint
failed. Under this understanding, the current behavior is fine. If we want to
change this behavior to only times we are past the window, we should remove all
instances of the metric incr except those inside the hintWindowExpired block.
Under the current patch, the metric will always be incremented if Hinted
Handoff is enabled and possibly again if hintWindowExpired or if hinted handoff
is disabled for the specific DC.
[~iamaleksey] - since you did the most work with hinted handoff most recently,
do you have an opinion on what this metric should reflect? Current
implementation of the incrPastWindow metric counts the times shouldHint fails
because of either hints being disabled or hint window expiration. This issue
could be used to change the metric to only be incremented when the hint window
has ended.
> Hint Metrics are updated even if hinted_hand-offs=false
> -------------------------------------------------------
>
> Key: CASSANDRA-11168
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11168
> Project: Cassandra
> Issue Type: Bug
> Reporter: Anubhav Kale
> Assignee: Anubhav Kale
> Priority: Minor
> Attachments: 0001-Hinted-Handoff-Fix.patch
>
>
> In our PROD logs, we noticed a lot of hint metrics even though we have
> disabled hinted handoffs.
> The reason is StorageProxy.ShouldHint has an inverted if condition.
> We should also wrap the if (hintWindowExpired) block in if
> (DatabaseDescriptor.hintedHandoffEnabled()).
> The fix is easy, and I can provide a patch.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
