rangareddy commented on issue #17121:
URL: https://github.com/apache/hudi/issues/17121#issuecomment-5366098544

   This issue was reviewed as part of the JIRA-migrated backlog triage 
(HUDI-9660).
   
   **Findings: still open.**
   
   The description identifies a real crash-consistency hole and states the fix 
clearly, so this is workable as written. Nothing on `master` indicates lockstep 
stepping was implemented - the handlers under 
`hudi-client/hudi-client-common/src/main/java/org/apache/hudi/table/upgrade/` 
are still per-version-pair (`SevenToEightUpgradeHandler`, 
`EightToSevenDowngradeHandler`, `SevenToSixDowngradeHandler` and so on), with 
the MDT and data table migrated in separate sweeps.
   
   Two related tickets worth linking, since all three touch the same 
sequencing: #17117 (HUDI-9651) asks whether to delete the MDT when it lags 
during a 7-to-8 upgrade - which the lockstep approach here would make 
unnecessary - and #17272 (HUDI-8237) asks for upgrade/downgrade to upgrade the 
MDT "in the same way", which sounds like the same intent stated differently.
   
   Recommend consolidating those three into one design, since implementing them 
independently would produce conflicting sequencing.
   
   Keeping this open.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to