[
https://issues.apache.org/jira/browse/KAFKA-866?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sriram Subramanian updated KAFKA-866:
-------------------------------------
Attachment: KAFKA-866.patch
1. Recovery decompresses the message to get the starting offset
2. AbstractFetcherThread should do deep iteration. This was missed in the
IndexOutOfBounds fix.
> Recover segment does shallow iteration to fix index causing inconsistencies
> ---------------------------------------------------------------------------
>
> Key: KAFKA-866
> URL: https://issues.apache.org/jira/browse/KAFKA-866
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.8
> Reporter: Sriram Subramanian
> Assignee: Sriram Subramanian
> Priority: Blocker
> Labels: kafka-0.8, p1
> Attachments: KAFKA-866.patch
>
>
> When we recover a segment, we do a shallow iteration to fix the index. This
> creates entries in the index with the last logical offset in a compressed
> messageSet. However, during log append we use the starting offset in a
> compressed message set to populate the index (We changed to starting offset
> recently). This causes inconsistencies.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
