[
https://issues.apache.org/jira/browse/OAK-5753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15882505#comment-15882505
]
Michael Dürig commented on OAK-5753:
------------------------------------
Looks good AFAICS.
Would it be helpful to also log the timestamp along with the revisions? This
would need an improvement for the {{JournalReader}} first, but since we now
have the timestamps in the journal that additional information might be
helpful. Especially in situations like this:
{noformat}
Latest good revision for path /x is b8428d1d-23d9-4995-a3c0-84b09ac93d8f:18
Latest good revision for path /y is a5a751db-3b0d-420e-ade4-34e83eb767cc:109
{noformat}
As here it would make clear which revision is actually older.
> Consistency check incorrectly fails for broken partial paths
> -------------------------------------------------------------
>
> Key: OAK-5753
> URL: https://issues.apache.org/jira/browse/OAK-5753
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: run, segment-tar
> Affects Versions: 1.8
> Reporter: Andrei Dulceanu
> Assignee: Andrei Dulceanu
> Labels: tooling
> Fix For: 1.7.0, 1.8
>
> Attachments: OAK-5753.patch
>
>
> To better explain the bug I'll describe the content of the revisions:
> # Valid Revision
> Adds child nodes {{a}}, {{b}}, {{c}}, {{d}}, {{e}}, {{f}} with various
> properties (blobs included)
> # Invalid Revision
> Adds child node {{z}} with some blob properties and then corrupts the
> {{NODE}} record holding {{z}}.
> Now when the consistency check is run, it correctly detects that the second
> revision is broken, *marks the path {{/z}} as corrupt* and then continues
> checking the first valid revision. Because of a check introduced for OAK-5556
> [1], which tries to validate the user provided absolute paths before checking
> them, the checker tries to check {{/z}} in the first revision, where of
> course it can't find it. Therefore the check incorrectly fails for this
> revision, although it shouldn't have to.
> /cc [~mduerig], [~frm]
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
