Github user mxm commented on a diff in the pull request: https://github.com/apache/flink/pull/934#discussion_r35635322 --- Diff: flink-runtime/src/main/java/org/apache/flink/runtime/executiongraph/ExecutionGraph.java --- @@ -619,6 +718,21 @@ public ExecutionContext getExecutionContext() { resultStrings[i++] = result; } + for (Map.Entry<String, List<BlobKey>> entry : largeAccumulatorMap.entrySet()) { + + if(!smallAccumulatorMap.containsKey(entry.getKey())) { + StringBuilder str = new StringBuilder(); + str.append("BlobKeys=[ "); + for (BlobKey bk : entry.getValue()) { + str.append(bk + " "); + } + str.append("]"); + + StringifiedAccumulatorResult result = + new StringifiedAccumulatorResult(entry.getKey(), "Blob/Serialized", str.toString()); + resultStrings[i++] = result; + } + } --- End diff -- Shouldn't this be handled transparently for the user? The user shouldn't have to care if the accumulators were received from the BlobManager or via an Akka message.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---