diff --git a/cluster/src/shard_operator.rs b/cluster/src/shard_operator.rs
index 129715b702..fab9a1add7 100644
--- a/cluster/src/shard_operator.rs
+++ b/cluster/src/shard_operator.rs
@@ -15,7 +15,6 @@ use snafu::ResultExt;
 use table_engine::{
     engine::{TableEngineRef, TableState},
     table::TableId,
-    ANALYTIC_ENGINE_TYPE,
 };
 
 use crate::{
diff --git a/cluster/src/shard_set.rs b/cluster/src/shard_set.rs
index a537964ffe..f1921900fb 100644
--- a/cluster/src/shard_set.rs
+++ b/cluster/src/shard_set.rs
@@ -13,8 +13,9 @@ use crate::{
     Result, ShardVersionMismatch, TableAlreadyExists, TableNotFound, UpdateFrozenShard,
 };
 
-/// [ShardTablesCache] caches the information about tables and shards, and the
-/// relationship between them is: one shard -> multiple tables.
+/// Shard set
+///
+/// Manage all shards opened on current node
 #[derive(Debug, Default, Clone)]
 pub struct ShardSet {
     inner: Arc<std::sync::RwLock<HashMap<ShardId, ShardRef>>>,
@@ -46,6 +47,9 @@ impl ShardSet {
     }
 }
 
+/// Shard
+///
+/// NOTICE: all write operations on a shard will be performed sequentially.
 pub struct Shard {
     pub data: ShardDataRef,
     pub operator: tokio::sync::Mutex<ShardOperator>,
@@ -113,12 +117,6 @@ impl Shard {
 
 pub type ShardRef = Arc<Shard>;
 
-#[derive(Debug, Clone)]
-pub struct TableWithShards {
-    pub table_info: TableInfo,
-    pub shard_infos: Vec<ShardInfo>,
-}
-
 #[derive(Debug, Clone)]
 pub struct UpdatedTableInfo {
     pub prev_version: u64,
@@ -126,6 +124,7 @@ pub struct UpdatedTableInfo {
     pub table_info: TableInfo,
 }
 
+/// Shard data
 #[derive(Debug)]
 pub struct ShardData {
     /// Shard info
diff --git a/server/src/grpc/meta_event_service/mod.rs b/server/src/grpc/meta_event_service/mod.rs
index f800d3f3d9..165f15b4c6 100644
--- a/server/src/grpc/meta_event_service/mod.rs
+++ b/server/src/grpc/meta_event_service/mod.rs
@@ -297,7 +297,7 @@ async fn do_open_shard(ctx: HandlerContext, shard_info: ShardInfo) -> Result<()>
 
     shard.open(open_ctx).await.box_err().context(ErrWithCause {
         code: StatusCode::Internal,
-        msg: "fail to open shard in cluster",
+        msg: "fail to open shard",
     })
 }
 
@@ -346,7 +346,7 @@ async fn do_close_shard(ctx: &HandlerContext, shard_id: ShardId) -> Result<()> {
         .box_err()
         .context(ErrWithCause {
             code: StatusCode::Internal,
-            msg: "fail to close shard in cluster",
+            msg: "fail to close shard",
         })?;
 
     // Remove the shard from the cluster topology after the shard is closed indeed.
@@ -434,7 +434,7 @@ async fn handle_create_table_on_shard(
         .box_err()
         .context(ErrWithCause {
             code: StatusCode::Internal,
-            msg: "fail to create table on shard in cluster",
+            msg: "fail to create table on shard",
         })
 }
 
@@ -470,7 +470,7 @@ async fn handle_drop_table_on_shard(
         .box_err()
         .context(ErrWithCause {
             code: StatusCode::Internal,
-            msg: "fail to drop table on shard in cluster",
+            msg: "fail to drop table on shard",
         })
 }
 
@@ -503,7 +503,7 @@ async fn handle_open_table_on_shard(
         .box_err()
         .context(ErrWithCause {
             code: StatusCode::Internal,
-            msg: "fail to open table on shard in cluster",
+            msg: "fail to open table on shard",
         })
 }
 
@@ -537,7 +537,7 @@ async fn handle_close_table_on_shard(
         .box_err()
         .context(ErrWithCause {
             code: StatusCode::Internal,
-            msg: "fail to close table on shard in cluster",
+            msg: "fail to close table on shard",
         })
 }