[ 
https://issues.apache.org/jira/browse/OAK-9841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17567997#comment-17567997
 ] 

Robert Munteanu commented on OAK-9841:
--------------------------------------

[~kwin] - I think the deployer is the one that should make this judgement. I 
wonder whether we can use the {{.target}} and {{.cardinality.minimum}} 
properties defined in 
https://docs.osgi.org/specification/osgi.cmpn/7.0.0/service.component.html#d0e40380
 for this.

> Race condition in NodeStoreChecksService
> ----------------------------------------
>
>                 Key: OAK-9841
>                 URL: https://issues.apache.org/jira/browse/OAK-9841
>             Project: Jackrabbit Oak
>          Issue Type: Bug
>          Components: composite
>            Reporter: Konrad Windszus
>            Priority: Major
>
> The {{NodeStoreChecksService}} 
> (https://github.com/apache/jackrabbit-oak/blob/bbc141fd1fb9ff0d9ce742279445df9eb698c3e3/oak-store-composite/src/main/java/org/apache/jackrabbit/oak/composite/checks/NodeStoreChecksService.java#L41)
>  executes all bound {{MountedNodeStoreChecker}} s which have been there at 
> the time of activation.
> The references are not greedily referenced and also there is no wait for 
> specific {{MountedNodeStoreChecker}} services to be active.
> That leads to the fact that the usage of {{NodeStoreChecksService}} in 
> {{CompositeNodeStoreService.registerCompositeNodeStore(...)}} is not 
> deterministic as the starting order of OSGi services determine which 
> {{MountedNodeStoreChecker}} are active during 
> {{CompositeNodeStoreService.activate()}}.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to