diff --git a/cassandra/src/main/scala/filodb.cassandra/metastore/CheckpointMigration.scala b/cassandra/src/main/scala/filodb.cassandra/metastore/CheckpointMigration.scala
new file mode 100644
index 0000000000..9b0a4d0c99
--- /dev/null
+++ b/cassandra/src/main/scala/filodb.cassandra/metastore/CheckpointMigration.scala
@@ -0,0 +1,135 @@
+package filodb.cassandra.metastore
+
+import java.lang.{Integer => JInt, Long => JLong}
+import com.datastax.driver.core.{ConsistencyLevel, Session}
+import com.typesafe.config.Config
+import com.typesafe.scalalogging.StrictLogging
+import filodb.cassandra.{FiloCassandraConnector, FiloSessionProvider}
+import filodb.core.memstore.FiloSchedulers
+import filodb.core.{DatasetRef, GlobalConfig}
+import monix.execution.{Scheduler, UncaughtExceptionReporter}
+
+import scala.Console.println
+import scala.concurrent.{ExecutionContext, Future}
+import scala.util.{Failure, Success}
+
+
+sealed class CheckpointTableConnector(val keyspace: String,
+                                      val config: Config,
+                                      val session: Session,
+                                      writeConsistencyLevel: ConsistencyLevel,
+                                      checkPointReadConsistencyLevel: ConsistencyLevel)
+                                      (implicit val ec: ExecutionContext) extends FiloCassandraConnector {
+  val tableString = s"${keyspace}.checkpoints"
+
+  val createCql =
+    s"""CREATE TABLE IF NOT EXISTS $tableString (
+       | databasename text,
+       | datasetname text,
+       | shardnum int,
+       | groupnum int,
+       | offset bigint,
+       | PRIMARY KEY ((databasename, datasetname, shardnum), groupnum)
+       |)""".stripMargin
+
+  lazy val readCheckpointCql =
+    session.prepare(
+      s"""SELECT groupnum, offset FROM $tableString WHERE
+         | databasename = ? AND
+         | datasetname = ? AND
+         | shardnum = ? """.stripMargin).setConsistencyLevel(checkPointReadConsistencyLevel)
+  // we want consistent reads during recovery
+
+  lazy val writeCheckpointCql = {
+    val statement = session.prepare(
+      s"""INSERT INTO $tableString (databasename, datasetname, shardnum, groupnum, offset)
+         | VALUES (?, ?, ?, ?, ?)""".stripMargin
+    )
+    statement.setConsistencyLevel(writeConsistencyLevel)
+    statement
+  }
+
+  import filodb.cassandra.Util._
+  import filodb.core._
+
+  def initialize(): Future[Response] = execCql(createCql)
+
+  def clearAll(): Future[Response] = execCql(s"TRUNCATE $tableString")
+
+  def writeCheckpoint(dataset: DatasetRef, shardNum: Int, groupNum: Int, offset: Long): Future[Response] = {
+    // TODO database name should not be an optional in internally since there is a default value. Punted for later.
+    execStmt(writeCheckpointCql.bind(dataset.database.getOrElse(""),
+      dataset.dataset, shardNum: JInt, groupNum: JInt, offset: JLong))
+  }
+
+  def readCheckpoints(dataset: DatasetRef, shardNum: Int): Future[Map[Int, Long]] = {
+    session.executeAsync(readCheckpointCql.bind(dataset.database.getOrElse(""),
+      dataset.dataset, shardNum: JInt))
+      .toIterator // future of Iterator
+      .map { it => it.map(r => r.getInt(0) -> r.getLong(1)).toMap }
+  }
+}
+
+
+object CheckpointMigration extends StrictLogging {
+
+  def main(rawArgs: Array[String]): Unit = {
+    val usage =
+    """
+      Usage:  [--shardnum num] [--preagg num] [--raw filename]
+    """
+
+    if (rawArgs.isEmpty || rawArgs.length != 6) {
+      println(usage)
+      System.exit(1)
+    }
+
+    val argMap = Map.newBuilder[String, Any]
+
+    rawArgs.sliding(2, 2).toList.collect {
+      case Array("--shardnum", shardnum: String) => argMap.+=("shardnum" -> shardnum.toInt)
+      case Array("--preagg", preagg: String) => argMap.+=("preagg" -> preagg)
+      case Array("--raw", raw: String) => argMap.+=("raw" -> raw)
+    }
+    val args = argMap.result()
+    println(args)
+
+
+    val allConfig = GlobalConfig.configToDisableAkkaCluster.withFallback(GlobalConfig.systemConfig)
+    val filoConfig = allConfig.getConfig("filodb")
+    val cassandraConfig = filoConfig.getConfig("cassandra")
+    val session = FiloSessionProvider.openSession(cassandraConfig)
+    val ingestionConsistencyLevel = ConsistencyLevel.valueOf(cassandraConfig.getString("ingestion-consistency-level"))
+    val checkpointReadConsistencyLevel = ConsistencyLevel.valueOf(
+       cassandraConfig.getString("checkpoint-read-consistency-level"))
+
+    val shardNum = args.get("shardnum").get.asInstanceOf[Int]
+    logger.info(s"filoConfig =${filoConfig}")
+
+    lazy val ioPool = Scheduler.io(name = FiloSchedulers.IOSchedName,
+        reporter = UncaughtExceptionReporter(
+          m => println(s"Uncaught Exception in FilodbCluster.ioPool =${m}")))
+    val rawMetaStore = new CheckpointTableConnector(
+      args.get("raw").get.asInstanceOf[String], cassandraConfig, session,
+      ingestionConsistencyLevel, checkpointReadConsistencyLevel)(ioPool)
+    val preaggMetaStore = new CheckpointTableConnector(args.get("preagg").get.asInstanceOf[String], cassandraConfig, session,
+      ingestionConsistencyLevel, checkpointReadConsistencyLevel)(ioPool)
+
+    val preaggDataRef = DatasetRef("prometheus_preagg")
+    for (shard <- 0 until shardNum) {
+      val read = preaggMetaStore.readCheckpoints(preaggDataRef, shard)
+      read.onComplete({
+        case Success(result) => {
+          println(s"result=${result}")
+          logger.info(s"result=${result}")
+          result.foreach {
+            case (group, offset) => {
+              rawMetaStore.writeCheckpoint(preaggDataRef, shard, group, offset)
+            }
+          }
+        }
+        case Failure(exception) => exception.printStackTrace()
+      })(ioPool)
+    }
+  }
+}