[ 
https://issues.apache.org/jira/browse/HDDS-17850?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Arafat Khan updated HDDS-17850:
-------------------------------
    Labels: newbie  (was: )

> Fix duplicate directory object ID in TestOmDBInsightEndPoint legacy bucket 
> test data
> ------------------------------------------------------------------------------------
>
>                 Key: HDDS-17850
>                 URL: https://issues.apache.org/jira/browse/HDDS-17850
>             Project: Apache Ozone
>          Issue Type: Test
>          Components: Ozone Recon
>            Reporter: Arafat Khan
>            Assignee: Arafat Khan
>            Priority: Major
>              Labels: newbie
>
> In TestOmDBInsightEndPoint#setUpOmData, the Legacy bucket directory 
> "dir4/dir7/" is written with
> the same name (DIR_FOUR) and object ID (DIR_FOUR_OBJECT_ID) as "dir4/". It 
> looks like a copy-paste
> mistake: the sibling entries use DIR_FIVE and DIR_SIX with their own IDs.
> The two directories share one object ID, so when the Legacy bucket is 
> processed as a file system
> bucket (ozone.om.enable.filesystem.paths=true), directory 21 gets itself as 
> its parent in the
> namespace summary. NSSummaryTaskDbEventHandler#propagateSizeUpwards then 
> recurses over the parent
> chain without end and fails with StackOverflowError. Every test in the class 
> fails in setUp.
> The test passes today only because ozone.om.enable.filesystem.paths defaults 
> to false, so this path
> is never reached. Setting the property to true, as other deployments do, 
> exposes it.
> Proposed change: write dir7 with its own name (DIR_SEVEN) and a new unique 
> object ID.
> Optional follow-up, as a separate issue: make propagateSizeUpwards stop and 
> log a warning when the
> parent chain loops, instead of overflowing the stack.
> Reproduce: set ozone.om.enable.filesystem.paths=true in the 
> OzoneConfiguration used by
> TestOmDBInsightEndPoint and run the class.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to