SOLR-17637: Fix LBHttpSolrClient & HttpShardHandler bug (#3147)

This bug causes async requests to be uncompleted in some error scenarios. The HttpShardHandler can hang indefinitely when this happens.
apache · Jan 30, 2025 · db0cbd3 · db0cbd3
1 parent 436e7f8
commit db0cbd3
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 2 deletions.
diff --git a/solr/CHANGES.txt b/solr/CHANGES.txt
@@ -182,6 +182,9 @@ Bug Fixes
   current live nodes list is stale. CloudSolrClient now retains the initial configured list of passed URLs as backup
   used for fetching cluster state when all live nodes have failed. (Matthew Biscocho via David Smiley, Houston Putman)
 
+* SOLR-17637: LBHttp2SolrClient can fail to complete async requests in certain error scenarios.
+  This can cause the HttpShardHandler to indefinitely wait on a completed response that will never come. (Houston Putman)
+
 Dependency Upgrades
 ---------------------
 * SOLR-17471: Upgrade Lucene to 9.12.1. (Pierre Salagnac, Christine Poerschke)

diff --git a/solr/solrj/src/java/org/apache/solr/client/solrj/impl/LBHttp2SolrClient.java b/solr/solrj/src/java/org/apache/solr/client/solrj/impl/LBHttp2SolrClient.java
@@ -156,7 +156,7 @@ public void onFailure(Exception e, boolean retryReq) {
               Endpoint url;
               try {
                 url = it.nextOrError(e);
-              } catch (SolrServerException ex) {
+              } catch (Throwable ex) {
                 apiFuture.completeExceptionally(e);
                 return;
               }

diff --git a/solr/solrj/src/java/org/apache/solr/client/solrj/impl/LBSolrClient.java b/solr/solrj/src/java/org/apache/solr/client/solrj/impl/LBSolrClient.java
@@ -315,7 +315,9 @@ public synchronized Endpoint nextOrError(Exception previousEx) throws SolrServer
       }
       // Skipping check time exceeded for the first request
       // Ugly string based hack but no live servers message here is VERY misleading :(
-      if ((previousEx != null && previousEx.getMessage().contains("Limits exceeded!"))
+      if ((previousEx != null
+              && previousEx.getMessage() != null
+              && previousEx.getMessage().contains("Limits exceeded!"))
           || (numServersTried > 0 && isTimeExceeded(timeAllowedNano, timeOutTime))) {
         throw new SolrServerException(
             "The processing limits for to this request were exceeded, see cause for details",