[
https://issues.apache.org/jira/browse/SLING-11856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18062508#comment-18062508
]
Carsten Ziegeler commented on SLING-11856:
------------------------------------------
Concurrency review finding potentially related to statistics issues:
BaseStatisticsImpl.add() uses nested synchronized locks: synchronized(this) {
synchronized(other) }. If a.add(b) and b.add(a) were ever called concurrently,
this would deadlock. While the current call flow is unidirectional (queue stats
-> global stats), the pattern is structurally unsafe and could contribute to
statistics inconsistencies. The existing copyFrom() method already provides a
safe alternative pattern: read other values into locals under
synchronized(other), then write under synchronized(this).
> Statistics for Sling Event Queue do not reset correctly
> -------------------------------------------------------
>
> Key: SLING-11856
> URL: https://issues.apache.org/jira/browse/SLING-11856
> Project: Sling
> Issue Type: Bug
> Components: Event
> Reporter: Roy Teeuwen
> Priority: Major
>
> When doing a reset on the statistics, the activeJobs and queuedJobs are not
> correctly reset.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
