andyhuangdev commented on code in PR #11650:
URL: https://github.com/apache/ozone/pull/11650#discussion_r4558976764


##########
hadoop-ozone/recon/src/main/java/org/apache/hadoop/ozone/recon/ReconServer.java:
##########
@@ -167,7 +171,15 @@ public Void call() throws Exception {
           injector.getInstance(ReconSchemaManager.class);
 
       LOG.info("Creating Recon Schema.");
-      reconSchemaManager.createReconSchema();
+      // A new SQL DB on an existing Recon (for example after the move from 
Derby to SQLite) cannot tell which upgrade
+      // actions the derived data already has, so start from the initial 
version to run all of them.
+      ReconVersion newSqlDbVersion = existingInstall ? 
ReconVersion.INITIAL_VERSION : ReconVersion.SOFTWARE_VERSION;
+      reconSchemaManager.createReconSchema(newSqlDbVersion);

Review Comment:
   Could we persist the pending NSSummary rebuild until it completes? The 
upgrade actions attempt to queue the rebuild before the OM metadata manager is 
initialized, so checkpoint creation fails and the retry is tracked only in 
memory. Finalization still advances the persisted schema version to the 
software version.
   
   If Recon restarts before the first sync, the retry flag is lost and the 
upgrade actions are skipped. The task cursors all default to zero in the new 
SQL DB, so the startup cursor comparison does not trigger a rebuild either.



-- 
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]


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

Reply via email to