Siyao Meng created HDDS-17628:
---------------------------------

             Summary: [Specula] Verify om-snapshot-diff-optimized (placeholder 
until HDDS-9154 is complete)
                 Key: HDDS-17628
                 URL: https://issues.apache.org/jira/browse/HDDS-17628
             Project: Apache Ozone
          Issue Type: Sub-task
            Reporter: Siyao Meng


Placeholder for a future Specula target. Do not run it yet.

This target will verify the optimized snapshot diff path from HDDS-9154: the 
key ordered merge of the two snapshots with a dependency ordered report 
(HDDS-15390, HDDS-15391, HDDS-15393, HDDS-15394), which HDDS-15397 put behind 
{{ozone.om.snapshot.diff.optimized.enabled}} (default false). Work under 
HDDS-9154 is still open: HDDS-15400 (enable it by default and deprecate the old 
path), HDDS-15401 (remove the flag), HDDS-15402 (remove the old code), 
HDDS-15399 and HDDS-16058. HDDS-16569 ({{om-snapshot-diff-semantics}}) ran 
before the optimized path was reachable and did not include it.

h3. When to revisit
Once the remaining HDDS-9154 work is resolved (at least HDDS-15400 and 
HDDS-15402): write the guidance at the master commit of that time, attach it, 
retitle this issue to the usual "[Specula] Verify om-snapshot-diff-optimized 
(Apache Ozone master @<sha>)" form, and add the help-wanted label. Until then 
the title does not follow that form on purpose, so tooling that picks the next 
open target skips it.

h3. Scope sketch
* Whether the report matches the legacy path for the same pair of snapshots, 
including renames, deletes of parents and children, and pagination.
* Whether applying the dependency ordered report in order to the older snapshot 
yields the newer one.
* Job submission, resubmission, cancellation, OM restart and leader change 
while a job runs.

Report equivalence alone may be better served by a differential property test 
against the legacy path; a model should focus on ordering and the job lifecycle.




--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to