example/cmd/microctl: Implement cluster recovery

Signed-off-by: Wesley Hershberger <[email protected]>

example/cmd/microctl/cluster_members.go

@@ -1,15 +1,46 @@
 package main
 
 import (
+	"bufio"
+	"fmt"
+	"io"
+	"os"
 	"sort"
+	"strings"
 
+	"github.com/canonical/lxd/shared"
 	cli "github.com/canonical/lxd/shared/cmd"
+	"github.com/canonical/lxd/shared/termios"
 	"github.com/spf13/cobra"
+	"golang.org/x/sys/unix"
+	"gopkg.in/yaml.v2"
 
 	"github.com/canonical/microcluster/client"
+	"github.com/canonical/microcluster/cluster"
 	"github.com/canonical/microcluster/microcluster"
 )
 
+const recoveryConfirmation = `You should only run this command if:
+ - A quorum of cluster members is permanently lost
+ - You are *absolutely* sure all microd instances are stopped
+ - This instance has the most up to date database
+
+Do you want to proceed? (yes/no): `
+
+const recoveryYamlComment = `# Member roles can be modified. Unrecoverable nodes should be given the role "spare".
+#
+# "voter" - Voting member of the database. A majority of voters is a quorum.
+# "stand-by" - Non-voting member of the database; can be promoted to voter.
+# "spare" - Not a member of the database.
+#
+# The edit is aborted if:
+# - the number of members changes
+# - the name of any member changes
+# - the ID of any member changes
+# - the address of any member changes
+# - no changes are made
+`
+
 type cmdClusterMembers struct {
 	common *CmdControl
 }
@@ -27,6 +58,9 @@ func (c *cmdClusterMembers) command() *cobra.Command {
 	var cmdList = cmdClusterMembersList{common: c.common}
 	cmd.AddCommand(cmdList.command())
 
+	var cmdRestore = cmdClusterEdit{common: c.common}
+	cmd.AddCommand(cmdRestore.command())
+
 	return cmd
 }
 
@@ -130,3 +164,73 @@ func (c *cmdClusterMemberRemove) run(cmd *cobra.Command, args []string) error {
 
 	return nil
 }
+
+type cmdClusterEdit struct {
+	common *CmdControl
+}
+
+func (c *cmdClusterEdit) command() *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "edit",
+		Short: "Recover the cluster from this node if quorum is lost",
+		RunE:  c.run,
+	}
+
+	return cmd
+}
+
+func (c *cmdClusterEdit) run(cmd *cobra.Command, args []string) error {
+	m, err := microcluster.App(microcluster.Args{StateDir: c.common.FlagStateDir, Verbose: c.common.FlagLogVerbose, Debug: c.common.FlagLogDebug})
+	if err != nil {
+		return err
+	}
+
+	members, err := m.GetLocalClusterMembers()
+	if err != nil {
+		return err
+	}
+
+	membersYaml, err := yaml.Marshal(members)
+	if err != nil {
+		return err
+	}
+
+	var content []byte
+	if !termios.IsTerminal(unix.Stdin) {
+		content, err = io.ReadAll(os.Stdin)
+		if err != nil {
+			return err
+		}
+	} else {
+		reader := bufio.NewReader(os.Stdin)
+		fmt.Print(recoveryConfirmation)
+
+		input, _ := reader.ReadString('\n')
+		input = strings.TrimSuffix(input, "\n")
+
+		if strings.ToLower(input) != "yes" {
+			fmt.Println("Cluster edit aborted; no changes made")
+			return nil
+		}
+
+		content, err = shared.TextEditor("", append([]byte(recoveryYamlComment), membersYaml...))
+		if err != nil {
+			return err
+		}
+	}
+
+	newMembers := []cluster.LocalMember{}
+	err = yaml.Unmarshal(content, &newMembers)
+	if err != nil {
+		return err
+	}
+
+	err = m.RecoverFromQuorumLoss(newMembers)
+	if err != nil {
+		return fmt.Errorf("cluster edit: %w", err)
+	}
+
+	fmt.Println("Cluster reconfigured successfully")
+
+	return nil
+}