[HDFS-2766] HA: test for case where standby partially reads log and then performs checkpoint

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: HA branch (HDFS-1623)
• Component/s: ha, namenode
• Target Version/s: HA branch (HDFS-1623)
• Hadoop Flags: Reviewed
• Assignee: Aaron Myers
• Reporter: Todd Lipcon
• Created: 07/Jan/12 00:29
• Updated: 11/Jan/12 12:54
• Resolved: 11/Jan/12 08:26

Description

Here's a potential bug case that we don't currently test for:

• SBN is reading a finalized edits file when NFS disappears halfway through (or some intermittent error happens)
• SBN performs a checkpoint and uploads it to the NN
• NN receives a checkpoint that doesn't correspond to the end of any log segment
• Both NN and SBN should be able to restart at this point.