[HBASE-17887] Row-level consistency is broken for read

• Type: Sub-task
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 1.3.0, 2.0.0
• Fix Version/s: 1.4.0, 1.3.2, 2.0.0
• Component/s: regionserver
• Hadoop Flags: Reviewed
• Release Note: Now we pass on list of memstoreScanners to the StoreScanner along with the new files to ensure that the StoreScanner sees the latest memstore after flush.
• Assignee: Chia-Ping Tsai
• Reporter: Umesh Agashe
• Created: 06/Apr/17 18:34
• Updated: 01/Jul/22 20:25
• Resolved: 12/May/17 11:47

Description

The scanner of latest memstore may be lost if we make quick flushes. The following step may help explain this issue.

1. put data_A (seq id = 10, active store data_A and snapshots is empty)
2. snapshot of 1st flush (active is empty and snapshot stores data_A)
3. put data_B (seq id = 11, active store data_B and snapshot store data_A)
4. create user scanner (read point = 11, so It should see the data_B)
5. commit of 1st flush
   • clear snapshot ((hfile_A has data_A, active store data_B, and snapshot is empty)
   • update the reader (the user scanner receives the hfile_A)
6. snapshot of 2st flush (active is empty and snapshot store data_B)
7. commit of 2st flush
   • clear snapshot (hfile_A has data_A, hfile_B has data_B, active is empty, and snapshot is empty) – this is critical piece.
   • update the reader (haven't happen)
8. user scanner update the kv scanners (it creates scanner of hfile_A but nothing of memstore)
9. user see the older data A – wrong result

Issue Links

• blocks: HBASE-17876 Release 1.2.6
• is related to: HBASE-14970 Backport HBASE-13082 and its sub-jira to branch-1
• relates to: HBASE-18145 The flush may cause the corrupt data for reading
• relates to: HBASE-17863 Procedure V2: Proc Executor cleanup. Split FINISHED state to two states: SUCCESS and FAILED.
• links to: RB