(accumulo) branch elasticity updated: Added check for tablets with errors, when looking for volume replacements. (#4287)

[kturner]

This is an automated email from the ASF dual-hosted git repository.

kturner pushed a commit to branch elasticity
in repository https://gitbox.apache.org/repos/asf/accumulo.git


The following commit(s) were added to refs/heads/elasticity by this push:
     new cca3683e5b Added check for tablets with errors, when looking for 
volume replacements. (#4287)
cca3683e5b is described below

commit cca3683e5b51f66787569e9b968b4599848e136e
Author: Arbaaz Khan <bazzy...@yahoo.com>
AuthorDate: Tue Feb 27 17:18:36 2024 -0500

    Added check for tablets with errors, when looking for volume replacements. 
(#4287)
    
    Added check for tablets with errors, when looking for volume replacements.
    
    fixes #4234
---
 .../org/apache/accumulo/manager/TabletGroupWatcher.java    | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git 
a/server/manager/src/main/java/org/apache/accumulo/manager/TabletGroupWatcher.java
 
b/server/manager/src/main/java/org/apache/accumulo/manager/TabletGroupWatcher.java
index e8bdd794a4..e7ea20413a 100644
--- 
a/server/manager/src/main/java/org/apache/accumulo/manager/TabletGroupWatcher.java
+++ 
b/server/manager/src/main/java/org/apache/accumulo/manager/TabletGroupWatcher.java
@@ -353,6 +353,7 @@ abstract class TabletGroupWatcher extends 
AccumuloDaemonThread {
     int[] counts = new int[TabletState.values().length];
     private int totalUnloaded;
     private long totalVolumeReplacements;
+    private int tabletsWithErrors;
   }
 
   private TableMgmtStats manageTablets(Iterator<TabletManagement> iter,
@@ -398,6 +399,7 @@ abstract class TabletGroupWatcher extends 
AccumuloDaemonThread {
             "Error on TabletServer trying to get Tablet management information 
for extent: {}. Error message: {}",
             tm.getExtent(), mtiError);
         this.metrics.incrementTabletGroupWatcherError(this.store.getLevel());
+        tableMgmtStats.tabletsWithErrors++;
         continue;
       }
 
@@ -670,7 +672,17 @@ abstract class TabletGroupWatcher extends 
AccumuloDaemonThread {
 
         iter = store.iterator(tableMgmtParams);
         var tabletMgmtStats = manageTablets(iter, tableMgmtParams, 
currentTServers, true);
-        lookForTabletsNeedingVolReplacement = 
tabletMgmtStats.totalVolumeReplacements != 0;
+
+        // If currently looking for volume replacements, determine if the next 
round needs to look.
+        if (lookForTabletsNeedingVolReplacement) {
+          // Continue to look for tablets needing volume replacement if there 
was an error
+          // processing tablets in the call to manageTablets() or if we are 
still performing volume
+          // replacement. We only want to stop looking for tablets that need 
volume replacement when
+          // we have successfully processed all tablet metadata and no more 
volume replacements are
+          // being performed.
+          lookForTabletsNeedingVolReplacement = 
tabletMgmtStats.totalVolumeReplacements != 0
+              || tabletMgmtStats.tabletsWithErrors != 0;
+        }
 
         // provide stats after flushing changes to avoid race conditions w/ 
delete table
         stats.end(managerState);