[
https://issues.apache.org/jira/browse/OAK-7720?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16598348#comment-16598348
]
Michael Dürig commented on OAK-7720:
------------------------------------
[~frm], I wouldn't worry about the verbosity. If we hit that code path
something is severely wrong and we need all the relevant data we can get.
Re. {{SegmentBufferWriter#dumpSegmentBuffer}}, I wonder whether and how much we
can reuse the code from {{Segment.toString}} here so we stick with a consistent
format. More importantly though, let's make the code logging this warning extra
resilient and wrap it into a {{try-catch}} block. After all we are in an
somewhat inconsistent state here and it would be a shame if we hit another
exception and that one would hide the initial one some hard to understand way.
> Log hex dump of too big segments
> --------------------------------
>
> Key: OAK-7720
> URL: https://issues.apache.org/jira/browse/OAK-7720
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: segment-tar
> Reporter: Francesco Mari
> Assignee: Francesco Mari
> Priority: Major
> Attachments: OAK-7720-01.patch
>
>
> {{SegmentBufferWriter}} throws an ISE when it detects that a too big segment
> is about to be flushed. The ISE contains some details about the segment, like
> the amount of referenced segments, the amount of records, they payload length
> and the total length, but they are insufficient. {{SegmentBufferWriter}}
> should print a hex dump of the segment to the logger in order to speed up the
> debugging of this error condition.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
