zstan commented on a change in pull request #9527:
URL: https://github.com/apache/ignite/pull/9527#discussion_r738274997



##########
File path: 
modules/core/src/main/java/org/apache/ignite/internal/processors/cluster/ClusterProcessor.java
##########
@@ -329,10 +339,51 @@ public void updateTag(String newTag) throws 
IgniteCheckedException {
      * </ul>
      */
     public void onLocalJoin() {
-        cluster.setId(locClusterId != null ? locClusterId : UUID.randomUUID());
+        Collection<ClusterNode> rmtNodes = cluster.forServers().nodes();
+        List<ClusterNode> rmtNodes0 = new ArrayList<>(rmtNodes);
+
+        rmtNodes0.sort(Comparator.comparing(ClusterNode::id));
+
+        @Nullable Collection<BaselineNode> bltNodes = 
cluster.currentBaselineTopology();
+
+        if (F.isEmpty(bltNodes)) {
+            log.info("Baseline node collection is empty.");
+
+            return;
+        }
+
+        @Nullable UUID first = null;
+
+        Collection<Object> srvIds = F.nodeConsistentIds(bltNodes);
+
+        for (ClusterNode node : rmtNodes0) {
+            if (F.contains(srvIds, node.consistentId())) {
+                first = node.id();
+
+                break;
+            }
+        }
+
+        ClusterNode locNode = ctx.config().getDiscoverySpi().getLocalNode();
+
+        if (first == locNode.id() || locClusterId != null) {
+            cluster.setId(locClusterId != null ? locClusterId : 
UUID.randomUUID());
+
+            cluster.setTag(locClusterTag != null ? locClusterTag :
+                ClusterTagGenerator.generateTag());
 
-        cluster.setTag(locClusterTag != null ? locClusterTag :
-            ClusterTagGenerator.generateTag());
+            ClusterIdAndTag idAndTag = new ClusterIdAndTag(cluster.id(), 
cluster.tag());
+
+            if (log.isInfoEnabled())
+                log.info("Writing cluster ID and tag to metastorage on ready 
for write " + idAndTag);
+
+            try {
+                metastorage.writeAsync(CLUSTER_ID_TAG_KEY, idAndTag);
+            }
+            catch (IgniteCheckedException e) {
+                ctx.failure().process(new 
FailureContext(FailureType.CRITICAL_ERROR, e));

Review comment:
       seems we can`t cal CAS when cluster is in transition state, like this 
case. I see that cas have no progress in this place.




-- 
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