[HBASE-13895] DATALOSS: Region assigned before WAL replay when abort

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.2.0
• Fix Version/s: 1.2.0, 1.1.2, 2.0.0
• Component/s: Recovery, Region Assignment, wal
• Hadoop Flags: Reviewed

Release Note

If the master went to assign a region concurrent with a RegionServer abort, the returned RegionServerAbortedException was being handled as though the region had been cleanly offlined so assign was allowed proceed. If the region was opened in its new location before WAL replay completion, the replayed edits were ignored, worst case, or were later played over the top of edits that had come in since open and so susceptible to overwrite. In either case, DATALOSS.

• Assignee: Michael Stack
• Reporter: Michael Stack
• Created: 12/Jun/15 21:37
• Updated: 24/Jun/22 19:20
• Resolved: 02/Jul/15 01:12

Description

Opening a place holder till finish analysis.

I have dataloss running ITBLL at 3B (testing HBASE-13877). Most obvious culprit is the double-assignment that I can see.

Sub-Tasks

• Retry when RegionServerNotYetRunningException rather than go ahead with assign so for sure we don't skip WAL replay: Closed