When deactivating a clickhouse cluster node, it is not possible to automatically connect to the normal node

Observed

I have built a clickhouse cluster with 2 nodes, using http protocol to connect to the cluster, when I stop a node, querying the data will report an error.

detail descption:

Use clickhouse-go to connect clickhouse cluster, cluster have 2 node(node1, node2)
When I stop clickhouse node1, The go program won't connect, it lasts for a minute and then it still won't connect
When I restart clickhouse node1, The go program recovery connect

Expected behaviour

It is possible to switch to the normal clickhouse node

Code example

// use http connect
func InitClickHouseClusterWithHttp(hosts []string, username, password, database string) (*sql.DB, error) {
    sqlDb := clickhouse.OpenDB(&clickhouse.Options{
        Addr: hosts,
        Auth: clickhouse.Auth{
            Database: database,
            Username: username,
            Password: password,
        },
        Protocol: clickhouse.HTTP,
    })
    return sqlDb, sqlDb.Ping()
}

// query data
func TestInitClickHouseClusterWithHttp(t *testing.T) {
    hosts := []string{
        "192.168.1.230:8128",
        "192.168.1.231:8129",
    }
    db, err := InitClickHouseClusterWithHttp(hosts, "username", "password", "tmp_db")
    if err != nil {
        fmt.Println(err)
    }

    for {
        rows, err := db.Query("SELECT uuid, ts, value FROM history_data order by ts limit 1")
        if err != nil {
            fmt.Println(err)
        }
        for rows.Next() {
            var (
                uuid  string
                ts    time.Time
                value float64
            )
            err = rows.Scan(&uuid, &ts, &value)
            if err != nil {
                fmt.Println(err)
            }
            fmt.Println(uuid, ts, value)
        }
        time.Sleep(1 * time.Second)
    }
}

Error log

Post "http://admin:***@192.168.1.230:8128?database=tmp_db&default_format=Native": dial tcp 192.168.1.230:8128: connectex: No connection could be made because the target machine actively refused it.

Details

Environment

[x] clickhouse-go: github.com/ClickHouse/clickhouse-go/v2 v2.26.0
[x] Interface: database/sql
[x] Go version: go1.21.6 windows/amd64
[x] Operating system: Windows11 23H2
[x] ClickHouse version: 24.3.2.23
[x] Is it a ClickHouse Cloud? no, build clickhouse cluster in local
[x] ClickHouse Server non-default settings, if any: use defualt setting

[x] CREATE TABLE statements for tables involved:

CREATE TABLE tmp_db.history_data
(
    `uuid` LowCardinality(String),
    `ts` DateTime CODEC(Delta(4),
 ZSTD(1)),
    `value` Float64 CODEC(Gorilla,
 ZSTD(1))
)
ENGINE = ReplicatedReplacingMergeTree('/clickhouse/tables/{shard}/tmp_history_data',
 '{replica}')
PARTITION BY toDate(ts)
ORDER BY (uuid,
 ts)
SETTINGS index_granularity = 8192;

[ ] Sample data for all these tables, use [clickhouse-obfuscator](https://github.com/ClickHouse/ClickHouse/blob/master/programs/obfuscator/Obfuscator.cpp#L42-L80) if necessary

ClickHouse / clickhouse-go

When deactivating a clickhouse cluster node, it is not possible to automatically connect to the normal node #1336