[
https://issues.apache.org/jira/browse/OAK-4740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15455750#comment-15455750
]
Francesco Mari commented on OAK-4740:
-------------------------------------
To solve this issue, I need to implement a {{TarRecovery}} strategy that parses
the raw segment data to extract the references to other segment and the
references to external binaries.
The references to other segments are easy to retrieve, since they are already
stored in the segment and easily accessible via the {{Segment}} API.
The references to external binaries are not as accessible, because those
references were removed for good when the unified index of references to
external binaries was introduced in the TAR file. To retrieve those references,
I need to parse the records stored in the segment and identify all of those
records that actually represent references to external binaries. This might not
be an easy task for a number of reasons. First, the lack of typing information
is a big issue when it comes to parse the data in the segment without resorting
to further context (see OAK-2498). Second, a reference to an external binary
might be too long to be stored in a data segment and should be read from a data
segment, further complicating the implementation of the {{TarRecovery}}. One
would argue why a binary ID should be bigger than ~16K, but this doesn't seem
to be the right place to rant about this.
I scheduled this issue a couple of versions ahead to have enough time to tackle
OAK-2498 and OAK-4659, whose combination might be beneficial to the resolution
of this issue.
> TarReader recovery skips generating the index and binary graphs
> ---------------------------------------------------------------
>
> Key: OAK-4740
> URL: https://issues.apache.org/jira/browse/OAK-4740
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: segment-tar
> Reporter: Alex Parvulescu
> Assignee: Francesco Mari
> Fix For: Segment Tar 0.0.16
>
>
> As noticed from the tar recovery bits [0] the resulting tar file would lack
> the binary reference graph and index graph. This has implications on the DSGC
> (not properly reporting binary references would result in binaries being
> GC'ed) and GC operations.
> / cc [~frm], [~mduerig]
> [0]
> https://github.com/apache/jackrabbit-oak/blob/trunk/oak-segment-tar/src/main/java/org/apache/jackrabbit/oak/segment/file/TarReader.java#L216
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
