[Enhancement] add more info in logging for the cases of exec plan fra…

…gment failed and clone (#17194)

Add related BE hostname to err message when query failed;
Add version info in be log when clone finished

Signed-off-by: Binglin Chang <decstery@gmail.com>
(cherry picked from commit 23e5e68)

be/src/storage/task/engine_clone_task.cpp

@@ -314,7 +314,8 @@ void EngineCloneTask::_set_tablet_info(Status status, bool is_new_tablet) {
         } else {
             LOG(INFO) << "clone get tablet info success. tablet_id:" << _clone_req.tablet_id
                       << ", schema_hash:" << _clone_req.schema_hash << ", signature:" << _signature
-                      << ", version:" << tablet_info.version;
+                      << ", version:" << tablet_info.version
+                      << ", min_readable_version:" << tablet_info.min_readable_version;
             _tablet_infos->push_back(tablet_info);
         }
     }

fe/fe-core/src/main/java/com/starrocks/clone/TabletSchedCtx.java

@@ -946,14 +946,15 @@ public void finishCloneTask(CloneTask cloneTask, TFinishTaskRequest request)
                 throw new SchedException(Status.UNRECOVERABLE, "tablet does not exist");
             }
 
+            String finishInfo = "";
             if (cloneTask.isLocal()) {
                 unprotectedFinishLocalMigration(request);
             } else {
-                unprotectedFinishClone(request, db, partition, replicationNum);
+                finishInfo = unprotectedFinishClone(request, db, partition, replicationNum);
             }
 
             state = State.FINISHED;
-            LOG.info("clone finished: {}", this);
+            LOG.info("clone finished: {} {}", this, finishInfo);
         } catch (SchedException e) {
             // if failed to too many times, remove this task
             ++failedRunningCounter;
@@ -988,8 +989,8 @@ private void unprotectedFinishLocalMigration(TFinishTaskRequest request) throws
         replica.updateVersion(reportedTablet.version);
     }
 
-    private void unprotectedFinishClone(TFinishTaskRequest request, Database db, Partition partition,
-                                        short replicationNum) throws SchedException {
+    private String unprotectedFinishClone(TFinishTaskRequest request, Database db, Partition partition,
+                                          short replicationNum) throws SchedException {
         List<Long> aliveBeIdsInCluster = infoService.getBackendIds(true);
         Pair<TabletStatus, TabletSchedCtx.Priority> pair = tablet.getHealthStatusWithPriority(
                 infoService, visibleVersion, replicationNum,
@@ -1055,6 +1056,8 @@ private void unprotectedFinishClone(TFinishTaskRequest request, Database db, Par
             // so we keep it state unchanged, and log update replica
             GlobalStateMgr.getCurrentState().getEditLog().logUpdateReplica(info);
         }
+        return String.format("version:%d min_readable_version:%d", reportedTablet.getVersion(),
+                reportedTablet.getMin_readable_version());
     }
 
     /*

fe/fe-core/src/main/java/com/starrocks/qe/Coordinator.java

@@ -1224,7 +1224,7 @@ private void deliverExecBatchFragmentsRequests(boolean enablePipelineEngine) thr
                             if (errMsg == null) {
                                 errMsg = "exec rpc error. backend id: " + pair.first.backend.getId();
                             }
-                            queryStatus.setStatus(errMsg);
+                            queryStatus.setStatus(errMsg + " backend:" + pair.first.address.hostname);
                             LOG.warn("exec plan fragment failed, errmsg={}, code: {}, fragmentId={}, backend={}:{}",
                                     errMsg, code, pair.first.fragmentId,
                                     pair.first.address.hostname, pair.first.address.port);