[8.19] Fix MergeWithLowDiskSpaceIT testRelocationWhileForceMerging (#132913)

* There must be at least one merge task blocked (#132496)

After triggering a force merge, there might still be other merge tasks
running concurrently, so it is incorrect to assert a single merge is
blocked in the test.

Fix #131789

Co-authored-by: Patrick Doyle <[email protected]>

* Use consistent terminology for transport version resources/references (#132882) (#132898)

The build system for transport versions operate on two different types
of projects. A project can have named references to transport versions,
and it could also be the source of truth for transport version resources,
where latest and definitions files exist. This commit renames the
plugins and tasks to consistently use that terminology.

---------

Co-authored-by: Patrick Doyle <[email protected]>
Co-authored-by: Ryan Ernst <[email protected]>

Author: 3 people

server/src/internalClusterTest/java/org/elasticsearch/index/engine/MergeWithLowDiskSpaceIT.java

@@ -179,6 +179,7 @@ public void testForceMergeIsBlockedThenUnblocked() throws Exception {
             indexName,
             Settings.builder().put(IndexMetadata.SETTING_NUMBER_OF_REPLICAS, 0).put(IndexMetadata.SETTING_NUMBER_OF_SHARDS, 1).build()
         );
+        ensureGreen(indexName);
         // get current disk space usage (for all indices on the node)
         IndicesStatsResponse stats = indicesAdmin().prepareStats().clear().setStore(true).get();
         long usedDiskSpaceAfterIndexing = stats.getTotal().getStore().sizeInBytes();
@@ -254,6 +255,7 @@ public void testRelocationWhileForceMerging() throws Exception {
             indexName,
             Settings.builder().put(IndexMetadata.SETTING_NUMBER_OF_REPLICAS, 0).put(IndexMetadata.SETTING_NUMBER_OF_SHARDS, 1).build()
         );
+        ensureGreen(indexName);
         // get current disk space usage (for all indices on the node)
         IndicesStatsResponse stats = indicesAdmin().prepareStats().clear().setStore(true).get();
         long usedDiskSpaceAfterIndexing = stats.getTotal().getStore().sizeInBytes();
@@ -287,8 +289,8 @@ public void testRelocationWhileForceMerging() throws Exception {
         ThreadPoolMergeExecutorService threadPoolMergeExecutorService = internalCluster().getInstance(IndicesService.class, node1)
             .getThreadPoolMergeExecutorService();
         assertBusy(() -> {
-            // merge executor says merging is blocked due to insufficient disk space while there is a single merge task enqueued
-            assertThat(threadPoolMergeExecutorService.getMergeTasksQueueLength(), equalTo(1));
+            // merge executor says merging is blocked due to insufficient disk space
+            assertThat(threadPoolMergeExecutorService.getMergeTasksQueueLength(), greaterThan(0));
             assertTrue(threadPoolMergeExecutorService.isMergingBlockedDueToInsufficientDiskSpace());
             // indices stats also says that no merge is currently running (blocked merges are NOT considered as "running")
             IndicesStatsResponse indicesStatsResponse = client().admin().indices().prepareStats(indexName).setMerge(true).get();