thangTang commented on a change in pull request #1071: HBASE-23693 Split
failure may cause region hole and data loss when use zk assign
URL: https://github.com/apache/hbase/pull/1071#discussion_r368631873
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/master/RegionStates.java
##########
@@ -783,10 +801,20 @@ public void regionOffline(
LOG.info("Found region in " + state +
" to be reassigned by ServerCrashProcedure for " + sn);
rits.add(hri);
- } else if(state.isSplittingNew() || state.isMergingNew()) {
- LOG.info("Offline/Cleanup region if no meta entry exists, hri: " +
hri +
- " state: " + state);
- regionsToClean.add(state.getRegion());
+ } else if (state.isSplittingNew() || state.isMergingNew()) {
+ LOG.info(
+ "Offline/Cleanup region if no meta entry exists, hri: " + hri +
" state: " + state);
+ if (daughter2Parent.containsKey(hri.getEncodedName())) {
+ HRegionInfo parent = daughter2Parent.get(hri.getEncodedName());
+ HRegionInfo info = getHRIFromMeta(parent);
+ if (info != null && info.isSplit() && info.isOffline()) {
+ regionsToClean.add(Pair.newPair(state.getRegion(), info));
Review comment:
If the RS machine crash when the SplitTransactionImpl step after PONR,
master will handle the split rollback. Under normal conditions, it cleanup
daughter region and try to assign parent region. If assign failed, parent
region RIT so CatalogJanitor blocked; but at this time the master switch, new
master will delete parent region because of **OFFLINE and SPLIT=true**.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
