[
https://issues.apache.org/jira/browse/CASSANDRA-18086?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17644439#comment-17644439
]
Ariel Weisberg edited comment on CASSANDRA-18086 at 12/7/22 5:25 PM:
---------------------------------------------------------------------
I am generally +1 on the fixes and tests just to unblock the release.
One line break was removed.
The test that actually checks to see if the wait times are correct would work
better if it mocked time and could also check an upper bound on the time that
was waited, but to unblock things I think it is fine. If there are multiple
code paths that calculate or implement backoff it would be good to iterate
through those configurations.
If {{testBackoffTimeElapsed}} were thorough enough we might not need
{{testBackoffTime}} since that is just an implementation detail.
was (Author: aweisberg):
I am generally +1 on the fixes and tests just to unblock the release.
One line break was removed.
The test that actually checks to see if the wait times are correct would work
better if it mocked time and could also check an upper bound on the time that
was waited, but to unblock things I think it is fine. If there are multiple
code paths that calculate or implement backoff it would be good to iterate
through those.
If {{testBackoffTimeElapsed}} were thorough enough we might not need
{{testBackoffTime}} since that is just an implementation detail.
> Bug fix for 'wait' time to be in nanoseconds for consistency instead of
> microseconds
> ------------------------------------------------------------------------------------
>
> Key: CASSANDRA-18086
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18086
> Project: Cassandra
> Issue Type: Bug
> Components: Feature/Lightweight Transactions
> Reporter: Marianne Lyne Manaog
> Assignee: Marianne Lyne Manaog
> Priority: Normal
> Fix For: 4.1.x, 4.x
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> While working on benchmarking Paxos improvements in OSS Cassandra,
> [~benedict] was asked to review the work and thus found that the wait time
> was input in microseconds but then output incorrectly in the
> ContentionStrategy.java file, i.e., by summing up a wait time in nanoseconds
> with another wait time in microseconds (two different units used mistakenly).
> So, Benedict suggested the fix whereby the output wait time was then enforced
> to be consistently in nanoseconds.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
