[chore](backup) limit the involved tablets in a backup job #39987 (#40080)

cherry pick from #39987

fe/fe-common/src/main/java/org/apache/doris/common/Config.java

@@ -1526,6 +1526,15 @@ public class Config extends ConfigBase {
     @ConfField(mutable = true, masterOnly = true)
     public static int max_backup_restore_job_num_per_db = 10;
 
+    /**
+     * Control the max num of tablets per backup job involved.
+     */
+    @ConfField(mutable = true, masterOnly = true, description = {
+        "用于控制每次 backup job 允许备份的 tablet 上限，以避免 OOM",
+        "Control the max num of tablets per backup job involved, to avoid OOM"
+    })
+    public static int max_backup_tablets_per_job = 300000;
+
     /**
      * whether to ignore table that not support type when backup, and not report exception.
      */

fe/fe-core/src/main/java/org/apache/doris/backup/BackupJob.java

@@ -513,6 +513,17 @@ public class BackupJob extends AbstractJob {
             }
         }
 
+        // Limit the max num of tablets involved in a backup job, to avoid OOM.
+        if (unfinishedTaskIds.size() > Config.max_backup_tablets_per_job) {
+            String msg = String.format("the num involved tablets %d exceeds the limit %d, "
+                    + "which might cause the FE OOM, change config `max_backup_tablets_per_job` "
+                    + "to change this limitation",
+                    unfinishedTaskIds.size(), Config.max_backup_tablets_per_job);
+            LOG.warn(msg);
+            status = new Status(ErrCode.COMMON_ERROR, msg);
+            return;
+        }
+
         backupMeta = new BackupMeta(copiedTables, copiedResources);
 
         // send tasks