Block RPC write operations on volume servers when maintenance mode is enabled (#8115)

* Boostrap persistent state for volume servers.

This PR implements logic load/save persistent state information for storages
associated with volume servers, and reporting state changes back to masters
via heartbeat messages.

More work ensues!

See https://github.com/seaweedfs/seaweedfs/issues/7977 for details.

* Block RPC operations writing to volume servers when maintenance mode is on.

weed/server/volume_server.go

@@ -1,6 +1,7 @@
 package weed_server
 
 import (
+	"fmt"
 	"net/http"
 	"sync"
 	"time"
@@ -171,3 +172,19 @@ func (vs *VolumeServer) Reload() {
 	v := util.GetViper()
 	vs.guard.UpdateWhiteList(append(vs.whiteList, util.StringSplit(v.GetString("guard.white_list"), ",")...))
 }
+
+// Returns whether a volume server is in maintenance (i.e. read-only) mode.
+func (vs *VolumeServer) MaintenanceMode() bool {
+	if vs.store == nil {
+		return false
+	}
+	return vs.store.State.Pb.GetMaintenance()
+}
+
+// Checks if a volume server is in maintenance mode, and returns an error explaining why.
+func (vs *VolumeServer) CheckMaintenanceMode() error {
+	if !vs.MaintenanceMode() {
+		return nil
+	}
+	return fmt.Errorf("volume server %s is in maintenance mode", vs.store.Id)
+}