Merge pull request 'jpictl: add --version and version command' (#44) from pr-amery-version into main

Reviewed-on: #44

.gitignore

@@ -1 +1,2 @@
 .tmp
+.version

.vscode/settings.json

@@ -1,9 +1,15 @@
 {
     "cSpell.words": [
+        "asciigoat",
         "ceph",
+        "cyclomatic",
         "darvaza",
         "gofrs",
         "jpictl",
+        "libdns",
+        "Lookuper",
+        "publicsuffix",
+        "Wrapf",
         "zerolog"
     ]
 }

cmd/jpictl/config.go

@@ -1,22 +1,82 @@
 package main
 
-import "git.jpi.io/amery/jpictl/pkg/cluster"
+import (
+	"os"
+
+	"darvaza.org/core"
+
+	"git.jpi.io/amery/jpictl/pkg/cluster"
+)
+
+const (
+	// DefaultConfigFile is read if -f/--config-file isn't specified.
+	// If it doesn't exist, m/ will be scanned
+	DefaultConfigFile = "cloud.yaml"
+
+	// DefaultClusterDir is the directory we will scan and write
+	// unless something else is indicated
+	DefaultClusterDir = "m"
+
+	// DefaultDomain indicates the domain to use unless
+	// something else is specified
+	DefaultDomain = "jpi.cloud"
+)
 
 // Config describes the repository
 type Config struct {
 	Base   string
 	Domain string
+
+	ConfigFile string
 }
 
+var forceScan bool
+
 var cfg = &Config{
-	Base:   "m",
-	Domain: "jpi.cloud",
+	Base:   DefaultClusterDir,
+	Domain: DefaultDomain,
 }
 
 // LoadZones loads all zones and machines in the config directory
+// or file
 func (cfg *Config) LoadZones(resolve bool) (*cluster.Cluster, error) {
+	var zones *cluster.Cluster
+	var err error
+
+	if !forceScan {
+		// try config file first
+		zones, err = cluster.NewFromConfig(cfg.ConfigFile,
+			cluster.ResolvePublicAddresses(resolve),
+			cluster.WithLogger(log),
+		)
+
+		switch {
+		case err == nil:
+			// file was good
+			return zones, nil
+		case !os.IsNotExist(err) || cfg.ConfigFile != DefaultConfigFile:
+			// file was bad
+			return nil, core.Wrap(err, "NewFromConfig(%q)", cfg.ConfigFile)
+		}
+	}
+
+	// default file doesn't exist. scan instead.
 	return cluster.NewFromDirectory(cfg.Base, cfg.Domain,
 		cluster.ResolvePublicAddresses(resolve),
 		cluster.WithLogger(log),
 	)
 }
+
+func init() {
+	flags := rootCmd.PersistentFlags()
+
+	flags.StringVarP(&cfg.Base, "scan-dir", "d",
+		DefaultClusterDir, "directory to scan for cluster data")
+	flags.StringVarP(&cfg.Domain, "domain", "D",
+		DefaultDomain, "domain to use for scanned data")
+	flags.StringVarP(&cfg.ConfigFile, "config-file", "f",
+		DefaultConfigFile, "config file (JSON or YAML)")
+
+	flags.BoolVarP(&forceScan, "force-scan", "S",
+		false, "ignore config file and scan the directory instead")
+}

cmd/jpictl/dns.go

@@ -0,0 +1,193 @@
+package main
+
+import (
+	"context"
+	"net/netip"
+	"os"
+	"time"
+
+	"darvaza.org/core"
+	"github.com/spf13/cobra"
+
+	"git.jpi.io/amery/jpictl/pkg/cluster"
+	"git.jpi.io/amery/jpictl/pkg/dns"
+)
+
+const (
+	// DNSSyncTimeout specifies how long are we willing to wait for a DNS
+	// synchronization
+	DNSSyncTimeout = 10 * time.Second
+)
+
+func newDNSManager(m *cluster.Cluster, provider dns.Provider) (*dns.Manager, error) {
+	domain := m.Domain
+	if m.Name != "" {
+		domain = m.Name + "." + domain
+	}
+
+	mgr, err := dns.NewManager(dns.WithDomain(domain), dns.WithLogger(log))
+	if err != nil {
+		return nil, err
+	}
+
+	if provider != nil {
+		// set provider only  if specified
+		err = dns.WithProvider(provider)(mgr)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	if err := populateDNSManager(mgr, m); err != nil {
+		return nil, err
+	}
+
+	return mgr, nil
+}
+
+func populateDNSManager(mgr *dns.Manager, m *cluster.Cluster) error {
+	var err error
+
+	ctx := context.TODO()
+
+	m.ForEachZone(func(z *cluster.Zone) bool {
+		z.ForEachMachine(func(p *cluster.Machine) bool {
+			err = mgr.AddHost(ctx, z.Name, p.ID, p.IsActive(), p.PublicAddresses...)
+			return err != nil
+		})
+
+		return err != nil
+	})
+	if err != nil {
+		return err
+	}
+
+	m.ForEachRegion(func(r *cluster.Region) bool {
+		r.ForEachZone(func(z *cluster.Zone) bool {
+			err = mgr.AddRegion(ctx, r.Name, z.Name)
+			return err != nil
+		})
+
+		return err != nil
+	})
+
+	return err
+}
+
+// revive:disable:flag-parameter
+func newDNSManagerCommand(_ *cobra.Command,
+	resolve bool, withCredentials bool) (*dns.Manager, error) {
+	// revive:enable:flag-parameter
+	var cred dns.Provider
+
+	if withCredentials {
+		var err error
+
+		cred, err = dns.DefaultDNSProvider()
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	m, err := cfg.LoadZones(resolve)
+	if err != nil {
+		return nil, err
+	}
+
+	return newDNSManager(m, cred)
+}
+
+// Command
+var dnsCmd = &cobra.Command{
+	Use: "dns",
+}
+
+var dnsWriteCmd = &cobra.Command{
+	Use:    "write",
+	Short:  "dns write generates public DNS records",
+	PreRun: setVerbosity,
+	RunE: func(cmd *cobra.Command, _ []string) error {
+		mgr, err := newDNSManagerCommand(cmd, true, false)
+		if err != nil {
+			return err
+		}
+
+		_, err = mgr.WriteTo(os.Stdout)
+		return err
+	},
+}
+
+var dnsSyncCmd = &cobra.Command{
+	Use:    "sync",
+	Short:  "dns sync updates public DNS records",
+	PreRun: setVerbosity,
+	RunE: func(cmd *cobra.Command, _ []string) error {
+		mgr, err := newDNSManagerCommand(cmd, true, true)
+		if err != nil {
+			return err
+		}
+
+		ctx, cancel := context.WithTimeout(context.Background(), DNSSyncTimeout)
+		defer cancel()
+
+		return mgr.Sync(ctx)
+	},
+}
+
+var dnsShowCmd = &cobra.Command{
+	Use:    "show [<name>...]",
+	Short:  "dns show lists entries on DNS for our domain",
+	PreRun: setVerbosity,
+	RunE: func(cmd *cobra.Command, args []string) error {
+		mgr, err := newDNSManagerCommand(cmd, true, true)
+		if err != nil {
+			return err
+		}
+
+		ctx, cancel := context.WithTimeout(context.Background(), DNSSyncTimeout)
+		defer cancel()
+
+		return mgr.Show(ctx, args...)
+	},
+}
+
+var dnsAddCmd = &cobra.Command{
+	Use:    "add <name> <address..>",
+	Short:  "dns add registers a new machine on the public DNS",
+	Args:   cobra.MinimumNArgs(2),
+	PreRun: setVerbosity,
+	RunE: func(cmd *cobra.Command, args []string) error {
+		var addrs []netip.Addr
+
+		for _, s := range args[1:] {
+			addr, err := core.ParseAddr(s)
+			switch {
+			case err != nil:
+				return core.Wrap(err, s)
+			case !addr.IsValid(), addr.IsUnspecified(), addr.IsPrivate(), addr.IsMulticast():
+				return core.Wrap(core.ErrInvalid, s)
+			default:
+				addrs = append(addrs, addr)
+			}
+		}
+
+		mgr, err := newDNSManagerCommand(cmd, true, true)
+		if err != nil {
+			return err
+		}
+
+		ctx, cancel := context.WithTimeout(context.Background(), DNSSyncTimeout)
+		defer cancel()
+
+		return mgr.Add(ctx, args[0], addrs...)
+	},
+}
+
+func init() {
+	rootCmd.AddCommand(dnsCmd)
+
+	dnsCmd.AddCommand(dnsWriteCmd)
+	dnsCmd.AddCommand(dnsSyncCmd)
+	dnsCmd.AddCommand(dnsShowCmd)
+	dnsCmd.AddCommand(dnsAddCmd)
+}

cmd/jpictl/log.go

@@ -3,9 +3,13 @@ package main
 import (
 	"fmt"
 
+	"darvaza.org/sidecar/pkg/logger/zerolog"
 	"darvaza.org/slog"
+	"github.com/spf13/cobra"
 )
 
+var log = zerolog.New(nil, slog.Error)
+
 // fatal is a convenience wrapper for slog.Logger.Fatal().Print()
 func fatal(err error, msg string, args ...any) {
 	l := log.Fatal()
@@ -19,3 +23,20 @@ func fatal(err error, msg string, args ...any) {
 
 	panic("unreachable")
 }
+
+var verbosity int
+
+// setVerbosity replaces the global logger using the
+// verbosity level specified via -v flags
+func setVerbosity(_ *cobra.Command, _ []string) {
+	desired := int8(slog.Error) + int8(verbosity)
+	if desired > 6 {
+		desired = 6
+	}
+	log = zerolog.New(nil, slog.LogLevel(desired))
+}
+
+func init() {
+	rootCmd.PersistentFlags().CountVarP(&verbosity, "verbosity", "v",
+		"increase the verbosity level to Warn, Info or Debug")
+}

cmd/jpictl/main.go

@@ -2,8 +2,8 @@
 package main
 
 import (
-	"darvaza.org/sidecar/pkg/logger/zerolog"
-	"darvaza.org/slog"
+	_ "embed"
+
 	"github.com/spf13/cobra"
 )
 
@@ -13,11 +13,10 @@ const (
 )
 
 var (
-	log       = zerolog.New(nil, slog.Error)
-	verbosity int
-	rootCmd   = &cobra.Command{
-		Use:   CmdName,
-		Short: "control tool for jpi.cloud",
+	rootCmd = &cobra.Command{
+		Use:     CmdName,
+		Short:   "control tool for jpi.cloud",
+		Version: version,
 	}
 )
 
@@ -26,16 +25,3 @@ func main() {
 		fatal(err, "")
 	}
 }
-
-func init() {
-	rootCmd.PersistentFlags().CountVarP(&verbosity, "verbosity", "v",
-		"increase the verbosity level to Warn, Info or Debug")
-}
-
-func setVerbosity(_ *cobra.Command, _ []string) {
-	desired := int8(slog.Error) + int8(verbosity)
-	if desired > 6 {
-		desired = 6
-	}
-	log = log.WithLevel(slog.LogLevel(desired))
-}

cmd/jpictl/version.go

@@ -0,0 +1,31 @@
+package main
+
+import (
+	_ "embed"
+
+	"fmt"
+	"os"
+
+	"github.com/spf13/cobra"
+)
+
+//go:generate sh -c "git describe | tr -d '\r\n' > .version"
+//go:embed .version
+var version string
+
+var versionCmd = &cobra.Command{
+	Use:   "version",
+	Short: "Returns jpictl's version",
+	Args:  cobra.NoArgs,
+	Run: func(_ *cobra.Command, _ []string) {
+		_, _ = fmt.Fprintf(os.Stdout, "%s\n", version)
+	},
+}
+
+func init() {
+	if version == "" {
+		version = "undetermined"
+	}
+
+	rootCmd.AddCommand(versionCmd)
+}

go.mod

@@ -4,44 +4,51 @@ go 1.19
 
 require (
 	asciigoat.org/ini v0.2.5
-	darvaza.org/core v0.9.8
-	darvaza.org/resolver v0.5.4
-	darvaza.org/sidecar v0.0.2
-	darvaza.org/slog v0.5.3
-	darvaza.org/slog/handlers/discard v0.4.5
+	darvaza.org/core v0.12.0
+	darvaza.org/resolver v0.9.2
+	darvaza.org/sidecar v0.4.0
+	darvaza.org/slog v0.5.7
+	darvaza.org/slog/handlers/discard v0.4.11
 	github.com/gofrs/uuid/v5 v5.0.0
 	github.com/hack-pad/hackpadfs v0.2.1
-	github.com/mgechev/revive v1.3.3
-	github.com/spf13/cobra v1.7.0
-	golang.org/x/crypto v0.12.0
+	github.com/libdns/cloudflare v0.1.0
+	github.com/libdns/libdns v0.2.1
+	github.com/mgechev/revive v1.3.7
+	github.com/spf13/cobra v1.8.0
+	golang.org/x/crypto v0.20.0
+	golang.org/x/net v0.21.0
 	gopkg.in/gcfg.v1 v1.2.3
 	gopkg.in/yaml.v3 v3.0.1
 )
 
 require (
 	asciigoat.org/core v0.3.9 // indirect
-	darvaza.org/slog/handlers/filter v0.4.5 // indirect
-	darvaza.org/slog/handlers/zerolog v0.4.5 // indirect
+	darvaza.org/cache/x/simplelru v0.1.8 // indirect
+	darvaza.org/slog/handlers/filter v0.4.9 // indirect
+	darvaza.org/slog/handlers/zerolog v0.4.9 // indirect
 	github.com/BurntSushi/toml v1.3.2 // indirect
 	github.com/chavacava/garif v0.1.0 // indirect
-	github.com/fatih/color v1.15.0 // indirect
+	github.com/fatih/color v1.16.0 // indirect
 	github.com/fatih/structtag v1.2.0 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
+	github.com/kr/pretty v0.3.1 // indirect
 	github.com/mattn/go-colorable v0.1.13 // indirect
-	github.com/mattn/go-isatty v0.0.19 // indirect
+	github.com/mattn/go-isatty v0.0.20 // indirect
 	github.com/mattn/go-runewidth v0.0.15 // indirect
 	github.com/mgechev/dots v0.0.0-20210922191527-e955255bf517 // indirect
-	github.com/miekg/dns v1.1.55 // indirect
+	github.com/miekg/dns v1.1.58 // indirect
 	github.com/mitchellh/go-homedir v1.1.0 // indirect
 	github.com/olekukonko/tablewriter v0.0.5 // indirect
 	github.com/pkg/errors v0.9.1 // indirect
-	github.com/rivo/uniseg v0.4.4 // indirect
-	github.com/rs/zerolog v1.30.0 // indirect
+	github.com/rivo/uniseg v0.4.7 // indirect
+	github.com/rogpeppe/go-internal v1.11.0 // indirect
+	github.com/rs/zerolog v1.32.0 // indirect
+	github.com/spf13/afero v1.11.0 // indirect
 	github.com/spf13/pflag v1.0.5 // indirect
-	golang.org/x/mod v0.12.0 // indirect
-	golang.org/x/net v0.14.0 // indirect
-	golang.org/x/sys v0.12.0 // indirect
-	golang.org/x/text v0.13.0 // indirect
-	golang.org/x/tools v0.12.0 // indirect
+	golang.org/x/mod v0.15.0 // indirect
+	golang.org/x/sync v0.6.0 // indirect
+	golang.org/x/sys v0.17.0 // indirect
+	golang.org/x/text v0.14.0 // indirect
+	golang.org/x/tools v0.18.0 // indirect
 	gopkg.in/warnings.v0 v0.1.2 // indirect
 )

go.sum

@@ -2,31 +2,34 @@ asciigoat.org/core v0.3.9 h1:hgDDz4ecm3ZvehX++m8A/IzAt+B5oDPiRtxatzfUHPQ=
 asciigoat.org/core v0.3.9/go.mod h1:CAaHwyw8MpAq4a1MYtN2dxJrsK+hmIdW50OndaQZYPI=
 asciigoat.org/ini v0.2.5 h1:4gRIp9rU+XQt8+HMqZO5R7GavMv9Yl2+N+je6djDIAE=
 asciigoat.org/ini v0.2.5/go.mod h1:gmXzJ9XFqf1NLk5nQkj04USQ4tMtdRJHNQX6vp3DzjU=
-darvaza.org/core v0.9.8 h1:luLxgfUc2pzuusYPo/Z/dC/qr9XZPKpSQw8/kS7zNUM=
-darvaza.org/core v0.9.8/go.mod h1:Dbme64naxeshQfxcVJX9ZT7AiGyIY8kldfuELVtf8mw=
-darvaza.org/resolver v0.5.4 h1:dlSBNV14yYsp7Kg7ipwYOMNsLbrpeXa8Z0HBTa0Ryxs=
-darvaza.org/resolver v0.5.4/go.mod h1:vHMkQUmHjaetFqG2ZLZJiQHsXEMGoTOFGm+NXwfndhE=
-darvaza.org/sidecar v0.0.2 h1:4H8FUxc43kkLjxdShN1CoxLTcoHQsZjDVwm7kt6eIK0=
-darvaza.org/sidecar v0.0.2/go.mod h1:yFC3Qt3j+uS7n9CMpLxwrA68z+FNJhENoenBc9zBJJo=
-darvaza.org/slog v0.5.3 h1:sQzmZXgqRh9oFMKBwEYrEpucLvKJVZxaxa2bHIA6GJ0=
-darvaza.org/slog v0.5.3/go.mod h1:59d+yi+C7gn4pDDuwbbOKawERpdXthFFk1Yc+Sv6XB0=
-darvaza.org/slog/handlers/discard v0.4.5 h1:RRykOItNolHyiUav57lG/GFBL33rcljoa0nWTpY+T0g=
-darvaza.org/slog/handlers/discard v0.4.5/go.mod h1:HYHfISQjMqcPbPoPZ92ib/u7s9JcXvF6OaygpPFwdF8=
-darvaza.org/slog/handlers/filter v0.4.5 h1:CX1bMzldd67e3y3s3Sh4jK8Lyo0WMvTGBB2lD315jhc=
-darvaza.org/slog/handlers/filter v0.4.5/go.mod h1:OuH9rHYg9CIErTJCZliMnFexBfP/HJ9PZ1V1VwSCZ1g=
-darvaza.org/slog/handlers/zerolog v0.4.5 h1:W4cgGORx4wImr+RL96CWSQGTdkZzKX6YHXPSYJvdoB4=
-darvaza.org/slog/handlers/zerolog v0.4.5/go.mod h1:mCoh/mIl8Nsa6Yu1Um7d7cos6RuEJzgaTXaX5LDRUao=
+darvaza.org/cache/x/simplelru v0.1.8 h1:rvFucut4wKYbsYc994yR3P0M08NqlsvZxr5G4QK82tw=
+darvaza.org/cache/x/simplelru v0.1.8/go.mod h1:Mv1isOJTcXYK+aK0AvUe+/3KpRTXDsYga6rdTS/upNs=
+darvaza.org/core v0.12.0 h1:LLtYh9RZJSd0sgPTDocofCc4H5jbutSUQgPbKt+8gcI=
+darvaza.org/core v0.12.0/go.mod h1:47Ydh67KnzjLNu1mzX3r2zpphbxQqEaihMsUq5GflQ4=
+darvaza.org/resolver v0.9.2 h1:sUX6LZ1eN5TzJW7L4m7HM+BvwBeWl8dYYDGVSe+AIhk=
+darvaza.org/resolver v0.9.2/go.mod h1:XWqPhrxoOKNzRuSozOwmE1M6QVqQL28jEdxylnIO8Nw=
+darvaza.org/sidecar v0.4.0 h1:wHghxzLsiT82WDBBUf34aTqtOvRBg4UbxVIJgKNXRVA=
+darvaza.org/sidecar v0.4.0/go.mod h1:fUzjcFM4rN3bSEl4BKvok3MLpZWEhEa9+0/egmtpfMY=
+darvaza.org/slog v0.5.7 h1:JWC0OqvzR435AidIRDp4T9kdWTURWkUjzP4R78Koq1Q=
+darvaza.org/slog v0.5.7/go.mod h1:12L03t+KYhsZ9IbfF+8if5w9Y91af2par+bSzeBVqIQ=
+darvaza.org/slog/handlers/discard v0.4.11 h1:wr34OnDoRaMV1eGgW7yUaupQxjkTnuHrJmYRPj64RHM=
+darvaza.org/slog/handlers/discard v0.4.11/go.mod h1:ynxyLmZzZ5mP4ACLhQs4MEuDyhkIzjz6DfBHUjhnIK4=
+darvaza.org/slog/handlers/filter v0.4.9 h1:xD8OBwlJytpiwTSDDZqUuNSOsJuaManXQiOj9WEStr8=
+darvaza.org/slog/handlers/filter v0.4.9/go.mod h1:t+sjcf1c46kAdf1TRiQmop91xlkteZrC4WDXoVwHgP8=
+darvaza.org/slog/handlers/zerolog v0.4.9 h1:08FjRnwRGtJsLLBnbgxVorb/bkgm5QEM/LXD2cxeCbM=
+darvaza.org/slog/handlers/zerolog v0.4.9/go.mod h1:PZYfx6eOxQfD+cXJQp52iwKgcD30QVYHoXxOCojAOdw=
 github.com/BurntSushi/toml v1.3.2 h1:o7IhLm0Msx3BaB+n3Ag7L8EVlByGnpq14C4YWiu/gL8=
 github.com/BurntSushi/toml v1.3.2/go.mod h1:CxXYINrC8qIiEnFrOxCa7Jy5BFHlXnUU2pbicEuybxQ=
 github.com/chavacava/garif v0.1.0 h1:2JHa3hbYf5D9dsgseMKAmc/MZ109otzgNFk5s87H9Pc=
 github.com/chavacava/garif v0.1.0/go.mod h1:XMyYCkEL58DF0oyW4qDjjnPWONs2HBqYKI+UIPD+Gww=
 github.com/coreos/go-systemd/v22 v22.5.0/go.mod h1:Y58oyj3AT4RCenI/lSvhwexgC+NSVTIJ3seZv2GcEnc=
-github.com/cpuguy83/go-md2man/v2 v2.0.2/go.mod h1:tgQtvFlXSQOSOSIRvRPT7W67SCa46tRHOmNcaadrF8o=
+github.com/cpuguy83/go-md2man/v2 v2.0.3/go.mod h1:tgQtvFlXSQOSOSIRvRPT7W67SCa46tRHOmNcaadrF8o=
+github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/fatih/color v1.15.0 h1:kOqh6YHBtK8aywxGerMG2Eq3H6Qgoqeo13Bk2Mv/nBs=
-github.com/fatih/color v1.15.0/go.mod h1:0h5ZqXfHYED7Bhv2ZJamyIOUej9KtShiJESRwBDUSsw=
+github.com/fatih/color v1.16.0 h1:zmkK9Ngbjj+K0yRhTVONQh1p/HknKYSlNT+vZCzyokM=
+github.com/fatih/color v1.16.0/go.mod h1:fL2Sau1YI5c0pdGEVCbKQbLXB6edEj1ZgiY4NijnWvE=
 github.com/fatih/structtag v1.2.0 h1:/OdNE99OxoI/PqaW/SuSK9uxxT3f/tcSZgon/ssNSx4=
 github.com/fatih/structtag v1.2.0/go.mod h1:mBJUNpUnHmRKrKlQQlmCrh5PuhftFbNv8Ys4/aAZl94=
 github.com/godbus/dbus/v5 v5.0.4/go.mod h1:xhWf0FNVPg57R7Z0UbKHbJfkEywrmjJnf7w5xrFpKfA=
@@ -36,39 +39,53 @@ github.com/hack-pad/hackpadfs v0.2.1 h1:FelFhIhv26gyjujoA/yeFO+6YGlqzmc9la/6iKMI
 github.com/hack-pad/hackpadfs v0.2.1/go.mod h1:khQBuCEwGXWakkmq8ZiFUvUZz84ZkJ2KNwKvChs4OrU=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
-github.com/mattn/go-colorable v0.1.12/go.mod h1:u5H1YNBxpqRaxsYJYSkiCWKzEfiAb1Gb520KVy5xxl4=
+github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
+github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
+github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
+github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
+github.com/libdns/cloudflare v0.1.0 h1:93WkJaGaiXCe353LHEP36kAWCUw0YjFqwhkBkU2/iic=
+github.com/libdns/cloudflare v0.1.0/go.mod h1:a44IP6J1YH6nvcNl1PverfJviADgXUnsozR3a7vBKN8=
+github.com/libdns/libdns v0.2.0/go.mod h1:yQCXzk1lEZmmCPa857bnk4TsOiqYasqpyOEeSObbb40=
+github.com/libdns/libdns v0.2.1 h1:Wu59T7wSHRgtA0cfxC+n1c/e+O3upJGWytknkmFEDis=
+github.com/libdns/libdns v0.2.1/go.mod h1:yQCXzk1lEZmmCPa857bnk4TsOiqYasqpyOEeSObbb40=
 github.com/mattn/go-colorable v0.1.13 h1:fFA4WZxdEF4tXPZVKMLwD8oUnCTTo08duU7wxecdEvA=
 github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovkB8vQcUbaXHg=
-github.com/mattn/go-isatty v0.0.14/go.mod h1:7GGIvUiUoEMVVmxf/4nioHXj79iQHKdU27kJ6hsGG94=
 github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
-github.com/mattn/go-isatty v0.0.19 h1:JITubQf0MOLdlGRuRq+jtsDlekdYPia9ZFsB8h/APPA=
 github.com/mattn/go-isatty v0.0.19/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
+github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
+github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
 github.com/mattn/go-runewidth v0.0.9/go.mod h1:H031xJmbD/WCDINGzjvQ9THkh0rPKHF+m2gUSrubnMI=
 github.com/mattn/go-runewidth v0.0.15 h1:UNAjwbU9l54TA3KzvqLGxwWjHmMgBUVhBiTjelZgg3U=
 github.com/mattn/go-runewidth v0.0.15/go.mod h1:Jdepj2loyihRzMpdS35Xk/zdY8IAYHsh153qUoGf23w=
 github.com/mgechev/dots v0.0.0-20210922191527-e955255bf517 h1:zpIH83+oKzcpryru8ceC6BxnoG8TBrhgAvRg8obzup0=
 github.com/mgechev/dots v0.0.0-20210922191527-e955255bf517/go.mod h1:KQ7+USdGKfpPjXk4Ga+5XxQM4Lm4e3gAogrreFAYpOg=
-github.com/mgechev/revive v1.3.3 h1:GUWzV3g185agbHN4ZdaQvR6zrLVYTUSA2ktvIinivK0=
-github.com/mgechev/revive v1.3.3/go.mod h1:NhpOtVtDbjYNDj697eDUBTobijCDHQKar4HDKc0TuTo=
-github.com/miekg/dns v1.1.55 h1:GoQ4hpsj0nFLYe+bWiCToyrBEJXkQfOOIvFGFy0lEgo=
-github.com/miekg/dns v1.1.55/go.mod h1:uInx36IzPl7FYnDcMeVWxj9byh7DutNykX4G9Sj60FY=
+github.com/mgechev/revive v1.3.7 h1:502QY0vQGe9KtYJ9FpxMz9rL+Fc/P13CI5POL4uHCcE=
+github.com/mgechev/revive v1.3.7/go.mod h1:RJ16jUbF0OWC3co/+XTxmFNgEpUPwnnA0BRllX2aDNA=
+github.com/miekg/dns v1.1.58 h1:ca2Hdkz+cDg/7eNF6V56jjzuZ4aCAE+DbVkILdQWG/4=
+github.com/miekg/dns v1.1.58/go.mod h1:Ypv+3b/KadlvW9vJfXOTf300O4UqaHFzFCuHz+rPkBY=
 github.com/mitchellh/go-homedir v1.1.0 h1:lukF9ziXFxDFPkA1vsr5zpc1XuPDn/wFntq5mG+4E0Y=
 github.com/mitchellh/go-homedir v1.1.0/go.mod h1:SfyaCUpYCn1Vlf4IUYiD9fPX4A5wJrkLzIz1N1q0pr0=
 github.com/olekukonko/tablewriter v0.0.5 h1:P2Ga83D34wi1o9J6Wh1mRuqd4mF/x/lgBS7N7AbDhec=
 github.com/olekukonko/tablewriter v0.0.5/go.mod h1:hPp6KlRPjbx+hW8ykQs1w3UBbZlj6HuIJcUGPhkA7kY=
+github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsKMEsNbx1VGcRFpLqf3715MtcvvzbA=
 github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
 github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/rivo/uniseg v0.2.0/go.mod h1:J6wj4VEh+S6ZtnVlnTBMWIodfgj8LQOQFoIToxlJtxc=
-github.com/rivo/uniseg v0.4.4 h1:8TfxU8dW6PdqD27gjM8MVNuicgxIjxpm4K7x4jp8sis=
-github.com/rivo/uniseg v0.4.4/go.mod h1:FN3SvrM+Zdj16jyLfmOkMNblXMcoc8DfTHruCPUcx88=
+github.com/rivo/uniseg v0.4.7 h1:WUdvkW8uEhrYfLC4ZzdpI2ztxP1I582+49Oc5Mq64VQ=
+github.com/rivo/uniseg v0.4.7/go.mod h1:FN3SvrM+Zdj16jyLfmOkMNblXMcoc8DfTHruCPUcx88=
+github.com/rogpeppe/go-internal v1.9.0/go.mod h1:WtVeX8xhTBvf0smdhujwtBcq4Qrzq/fJaraNFVN+nFs=
+github.com/rogpeppe/go-internal v1.11.0 h1:cWPaGQEPrBb5/AsnsZesgZZ9yb1OQ+GOISoDNXVBh4M=
+github.com/rogpeppe/go-internal v1.11.0/go.mod h1:ddIwULY96R17DhadqLgMfk9H9tvdUzkipdSkR5nkCZA=
 github.com/rs/xid v1.5.0/go.mod h1:trrq9SKmegXys3aeAKXMUTdJsYXVwGY3RLcfgqegfbg=
-github.com/rs/zerolog v1.30.0 h1:SymVODrcRsaRaSInD9yQtKbtWqwsfoPcRff/oRXLj4c=
-github.com/rs/zerolog v1.30.0/go.mod h1:/tk+P47gFdPXq4QYjvCmT5/Gsug2nagsFWBWhAiSi1w=
+github.com/rs/zerolog v1.32.0 h1:keLypqrlIjaFsbmJOBdB/qvyF8KEtCWHwobLp5l/mQ0=
+github.com/rs/zerolog v1.32.0/go.mod h1:/7mN4D5sKwJLZQ2b/znpjC3/GQWY/xaDXUM0kKWRHss=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
-github.com/spf13/cobra v1.7.0 h1:hyqWnYt1ZQShIddO5kBpj3vu05/++x6tJ6dg8EC572I=
-github.com/spf13/cobra v1.7.0/go.mod h1:uLxZILRyS/50WlhOIKD7W6V5bgeIt+4sICxh6uRMrb0=
+github.com/spf13/afero v1.11.0 h1:WJQKhtpdm3v2IzqG8VMqrr6Rf3UYpEF239Jy9wNepM8=
+github.com/spf13/afero v1.11.0/go.mod h1:GH9Y3pIexgf1MTIWtNGyogA5MwRIDXGUr+hbWNoBjkY=
+github.com/spf13/cobra v1.8.0 h1:7aJaZx1B85qltLMc546zn58BxxfZdR/W22ej9CFoEf0=
+github.com/spf13/cobra v1.8.0/go.mod h1:WXLWApfZ71AjXPya3WOlMsY9yMs7YeiHhFVlvLyhcho=
 github.com/spf13/pflag v1.0.5 h1:iy+VFUOCP1a+8yFto/drg2CJ5u0yRoB7fZw3DKv/JXA=
 github.com/spf13/pflag v1.0.5/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
@@ -78,26 +95,26 @@ github.com/stretchr/testify v1.7.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/
 github.com/stretchr/testify v1.8.0/go.mod h1:yNjHg4UonilssWZ8iaSj1OCr/vHnekPRkoO+kdMU+MU=
 github.com/stretchr/testify v1.8.4 h1:CcVxjf3Q8PM0mHUKJCdn+eZZtm5yQwehR5yeSVQQcUk=
 github.com/stretchr/testify v1.8.4/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
-golang.org/x/crypto v0.12.0 h1:tFM/ta59kqch6LlvYnPa0yx5a83cL2nHflFhYKvv9Yk=
-golang.org/x/crypto v0.12.0/go.mod h1:NF0Gs7EO5K4qLn+Ylc+fih8BSTeIjAP05siRnAh98yw=
-golang.org/x/exp v0.0.0-20230713183714-613f0c0eb8a1 h1:MGwJjxBy0HJshjDNfLsYO8xppfqWlA5ZT9OhtUUhTNw=
-golang.org/x/mod v0.12.0 h1:rmsUpXtvNzj340zd98LZ4KntptpfRHwpFOHG188oHXc=
-golang.org/x/mod v0.12.0/go.mod h1:iBbtSCu2XBx23ZKBPSOrRkjjQPZFPuis4dIYUhu/chs=
-golang.org/x/net v0.14.0 h1:BONx9s002vGdD9umnlX1Po8vOZmrgH34qlHcD1MfK14=
-golang.org/x/net v0.14.0/go.mod h1:PpSgVXXLK0OxS0F31C1/tv6XNguvCrnXIDrFMspZIUI=
-golang.org/x/sync v0.3.0 h1:ftCYgMx6zT/asHUrPw8BLLscYtGznsLAnjq5RH9P66E=
-golang.org/x/sys v0.0.0-20210630005230-0f9fa26af87c/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210927094055-39ccf1dd6fa6/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/crypto v0.20.0 h1:jmAMJJZXr5KiCw05dfYK9QnqaqKLYXijU23lsEdcQqg=
+golang.org/x/crypto v0.20.0/go.mod h1:Xwo95rrVNIoSMx9wa1JroENMToLWn3RNVrTBpLHgZPQ=
+golang.org/x/exp v0.0.0-20240119083558-1b970713d09a h1:Q8/wZp0KX97QFTc2ywcOE0YRjZPVIx+MXInMzdvQqcA=
+golang.org/x/mod v0.15.0 h1:SernR4v+D55NyBH2QiEQrlBAnj1ECL6AGrA5+dPaMY8=
+golang.org/x/mod v0.15.0/go.mod h1:hTbmBsO62+eylJbnUtE2MGJUyE7QWk4xUqPFrRgJ+7c=
+golang.org/x/net v0.21.0 h1:AQyQV4dYCvJ7vGmJyKki9+PBdyvhkSd8EIx/qb0AYv4=
+golang.org/x/net v0.21.0/go.mod h1:bIjVDfnllIU7BJ2DNgfnXvpSvtn8VRwhlsaeUTyUS44=
+golang.org/x/sync v0.6.0 h1:5BMeUDZ7vkXGfEr1x9B4bRcTH4lpkTkpdh0T/J+qjbQ=
+golang.org/x/sync v0.6.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
 golang.org/x/sys v0.0.0-20220811171246-fbc7d0a398ab/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.12.0 h1:CM0HF96J0hcLAwsHPJZjfdNzs0gftsLfgKt57wWHJ0o=
 golang.org/x/sys v0.12.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/text v0.13.0 h1:ablQoSUd0tRdKxZewP80B+BaqeKJuVhuRxj/dkrun3k=
-golang.org/x/text v0.13.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
-golang.org/x/tools v0.12.0 h1:YW6HUoUmYBpwSgyaGaZq1fHjrBjX1rlpZ54T6mu2kss=
-golang.org/x/tools v0.12.0/go.mod h1:Sc0INKfu04TlqNoRA1hgpFZbhYXHPr4V5DzpSBTPqQM=
-gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
+golang.org/x/sys v0.17.0 h1:25cE3gD+tdBA7lp7QfhuV+rJiE9YXTcS3VG1SqssI/Y=
+golang.org/x/sys v0.17.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/text v0.14.0 h1:ScX5w1eTa3QqT8oi6+ziP7dTV1S2+ALU0bI+0zXKWiQ=
+golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
+golang.org/x/tools v0.18.0 h1:k8NLag8AGHnn+PHbl7g43CtqZAwG60vZkLqgyZgIHgQ=
+golang.org/x/tools v0.18.0/go.mod h1:GL7B4CwcLLeo59yx/9UWWuNOW1n3VZ4f5axWfML7Lcg=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/gcfg.v1 v1.2.3 h1:m8OOJ4ccYHnx2f4gQwpno8nAX5OGOh7RLaaz0pj3Ogs=
 gopkg.in/gcfg.v1 v1.2.3/go.mod h1:yesOnuUOFQAhST5vPY4nbZsb/huCgGGXlipJsBn0b3o=
 gopkg.in/warnings.v0 v0.1.2 h1:wFXVbFY8DY5/xOe1ECiWdKCzZlxgshcYVNkBHstARME=

pkg/ceph/config.go

@@ -39,6 +39,9 @@ func writeGlobalToBuffer(w *bytes.Buffer, c *GlobalConfig) {
 	_, _ = fmt.Fprintf(w, "%s = %s\n", "mon_initial_members", strings.Join(c.Monitors, ", "))
 	_, _ = fmt.Fprintf(w, "%s = %s\n", "mon_host", joinAddrs(c.MonitorsAddr, ", "))
 	_, _ = fmt.Fprintf(w, "%s = %s\n", "cluster_network", c.ClusterNetwork.String())
+
+	_, _ = fmt.Fprintf(w, "\n; %s\n", "don't rewrite labels on startup")
+	_, _ = fmt.Fprintf(w, "%s = %s\n", "osd_class_update_on_start", "false")
 }
 
 func joinAddrs(addrs []netip.Addr, sep string) string {

pkg/ceph/config_parser.go

@@ -17,7 +17,7 @@ var sectionMap = map[string]func(*Config, *basic.Section) error{
 func loadConfSection(out *Config, src *basic.Section) error {
 	h, ok := sectionMap[src.Key]
 	if !ok {
-		return core.Wrapf(fs.ErrInvalid, "unknown section %q", src.Key)
+		return core.Wrap(fs.ErrInvalid, "unknown section %q", src.Key)
 	}
 
 	return h(out, src)
@@ -48,7 +48,7 @@ func loadGlobalConfField(cfg *GlobalConfig, field basic.Field) error {
 	switch field.Key {
 	case "fsid":
 		if !core.IsZero(cfg.FSID) {
-			return core.Wrapf(fs.ErrInvalid, "duplicate field %q", field.Key)
+			return core.Wrap(fs.ErrInvalid, "duplicate field %q", field.Key)
 		}
 
 		err := cfg.FSID.UnmarshalText([]byte(field.Value))

pkg/cluster/ceph_scan.go

@@ -71,7 +71,7 @@ func newCephScanTODO(cfg *ceph.Config) *cephScanTODO {
 	return todo
 }
 
-func (m *Cluster) scanCephMonitors(_ *ScanOptions) error {
+func (m *Cluster) scanCephMonitors(opts *ScanOptions) error {
 	cfg, err := m.GetCephConfig()
 	switch {
 	case os.IsNotExist(err):
@@ -94,6 +94,10 @@ func (m *Cluster) scanCephMonitors(_ *ScanOptions) error {
 		todo.LogMissing(m.log)
 	}
 
+	return m.initCephMonitors(opts)
+}
+
+func (m *Cluster) initCephMonitors(_ *ScanOptions) error {
 	// make sure every zone has one
 	m.ForEachZone(func(z *Zone) bool {
 		_ = z.GetCephMonitors()

pkg/cluster/cluster.go

@@ -27,7 +27,8 @@ type Cluster struct {
 	Domain  string `json:"domain,omitempty" yaml:"domain,omitempty"`
 
 	CephFSID uuid.UUID `json:"ceph_fsid,omitempty" yaml:"ceph_fsid,omitempty"`
-	Zones    []*Zone   `json:"zones,omitempty"     yaml:"zones,omitempty"`
+	Regions  []Region  `json:",omitempty"          yaml:",omitempty"`
+	Zones    []*Zone   `json:",omitempty"          yaml:",omitempty"`
 }
 
 // revive:enable:line-length-limit

pkg/cluster/cluster_file.go

@@ -1,9 +1,12 @@
 package cluster
 
 import (
+	"bufio"
+	"bytes"
 	"fmt"
 	"io"
 	"os"
+	"strings"
 
 	fs "github.com/hack-pad/hackpadfs"
 )
@@ -33,7 +36,26 @@ func (m *Cluster) openWriter(name string, flags int, args ...any) (io.WriteClose
 		return nil, err
 	}
 
-	return f.(io.WriteCloser), nil
+	if f, ok := f.(io.WriteCloser); ok {
+		return f, nil
+	}
+
+	panic("unreachable")
+}
+
+// RemoveFile deletes a file from the cluster's config directory
+func (m *Cluster) RemoveFile(name string, args ...any) error {
+	if len(args) > 0 {
+		name = fmt.Sprintf(name, args...)
+	}
+
+	err := fs.Remove(m.dir, name)
+	switch {
+	case os.IsNotExist(err):
+		return nil
+	default:
+		return err
+	}
 }
 
 // ReadFile reads a file from the cluster's config directory
@@ -44,3 +66,50 @@ func (m *Cluster) ReadFile(name string, args ...any) ([]byte, error) {
 
 	return fs.ReadFile(m.dir, name)
 }
+
+// ReadLines reads a file from the cluster's config directory,
+// split by lines, trimmed, and accepting `#` to comment lines out.
+func (m *Cluster) ReadLines(name string, args ...any) ([]string, error) {
+	var out []string
+
+	data, err := m.ReadFile(name, args...)
+	if err != nil {
+		return nil, err
+	}
+
+	sc := bufio.NewScanner(bytes.NewReader(data))
+	for sc.Scan() {
+		s := strings.TrimSpace(sc.Text())
+		switch {
+		case s == "", strings.HasPrefix(s, "#"):
+			// ignore
+		default:
+			// accepted
+			out = append(out, s)
+		}
+	}
+
+	return out, nil
+}
+
+// WriteStringFile writes the given content to a file on the machine's config directory
+func (m *Cluster) WriteStringFile(value string, name string, args ...any) error {
+	f, err := m.CreateTruncFile(name, args...)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	buf := bytes.NewBufferString(value)
+	_, err = buf.WriteTo(f)
+	return err
+}
+
+// MkdirAll creates directories relative to the cluster's config directory
+func (m *Cluster) MkdirAll(name string, args ...any) error {
+	if len(args) > 0 {
+		name = fmt.Sprintf(name, args...)
+	}
+
+	return fs.MkdirAll(m.dir, name, 0755)
+}

pkg/cluster/cluster_fs.go

@@ -0,0 +1,25 @@
+package cluster
+
+import (
+	"io/fs"
+	"path/filepath"
+
+	"github.com/hack-pad/hackpadfs/os"
+)
+
+// DirFS returns a file system (an [fs.FS]) for the tree
+// of files rooted at the directory dir.
+func DirFS(dir string) (fs.FS, error) {
+	dir = filepath.Clean(dir)
+	fullPath, err := filepath.Abs(dir)
+	if err != nil {
+		return nil, err
+	}
+
+	sub, err := os.NewFS().Sub(fullPath[1:])
+	if err != nil {
+		return nil, err
+	}
+
+	return sub, nil
+}

pkg/cluster/cluster_import.go

@@ -0,0 +1,139 @@
+package cluster
+
+import (
+	"encoding/json"
+	"fmt"
+	"os"
+
+	"gopkg.in/yaml.v3"
+)
+
+func (m *Cluster) init(opts *ScanOptions) error {
+	for _, fn := range []func(*ScanOptions) error{
+		m.initZones,
+		m.scanZoneIDs,
+		m.scanSort,
+		m.scanGateways,
+		m.initCephMonitors,
+		m.initRegions,
+	} {
+		if err := fn(opts); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (m *Cluster) initZones(opts *ScanOptions) error {
+	var err error
+
+	sub, err := DirFS(m.BaseDir)
+	if err != nil {
+		return err
+	}
+
+	m.dir = sub
+
+	m.ForEachZone(func(z *Zone) bool {
+		err = m.initZone(z, opts)
+		return err != nil
+	})
+
+	return err
+}
+
+func (m *Cluster) initZone(z *Zone, _ *ScanOptions) error {
+	var hasMissing bool
+	var lastMachineID int
+
+	z.zones = m
+	z.logger = m
+
+	z.ForEachMachine(func(p *Machine) bool {
+		p.zone = z
+		p.logger = z
+
+		switch {
+		case p.ID == 0:
+			hasMissing = true
+		case p.ID > lastMachineID:
+			lastMachineID = z.ID
+		}
+
+		return false
+	})
+
+	if hasMissing {
+		next := lastMachineID + 1
+
+		z.ForEachMachine(func(p *Machine) bool {
+			if p.ID == 0 {
+				p.ID, next = next, next+1
+			}
+
+			return false
+		})
+	}
+
+	z.ForEachMachine(func(p *Machine) bool {
+		p.Name = fmt.Sprintf("%s-%v", z.Name, p.ID)
+		return false
+	})
+
+	return nil
+}
+
+func decodeConfigData(data []byte) (out *Cluster, err error) {
+	// try JSON first
+	out = new(Cluster)
+	err = json.Unmarshal(data, out)
+	if err == nil {
+		// good json
+		return out, nil
+	} else if _, ok := err.(*json.SyntaxError); !ok {
+		// bad json
+		return nil, err
+	}
+
+	out = new(Cluster)
+	err = yaml.Unmarshal(data, out)
+	if err != nil {
+		// bad yaml too
+		return nil, err
+	}
+
+	// good yaml
+	return out, nil
+}
+
+// NewFromConfig loads the cluster data from the given file
+func NewFromConfig(filename string, opts ...ScanOption) (*Cluster, error) {
+	var scanOptions ScanOptions
+
+	data, err := os.ReadFile(filename)
+	if err != nil {
+		return nil, err
+	}
+
+	m, err := decodeConfigData(data)
+	if err != nil {
+		return nil, err
+	}
+
+	for _, opt := range opts {
+		if err = opt(m, &scanOptions); err != nil {
+			return nil, err
+		}
+	}
+
+	if err = m.setScanDefaults(&scanOptions); err != nil {
+		return nil, err
+	}
+
+	if err := m.init(&scanOptions); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}

pkg/cluster/cluster_scan.go

@@ -2,17 +2,25 @@ package cluster
 
 import (
 	"io/fs"
+	"path"
 	"sort"
 
 	"darvaza.org/core"
 )
 
+const (
+	// ZoneRegionsFileName indicates the file containing
+	// region names as references
+	ZoneRegionsFileName = "regions"
+)
+
 func (m *Cluster) scan(opts *ScanOptions) error {
 	for _, fn := range []func(*ScanOptions) error{
 		m.scanDirectory,
 		m.scanMachines,
 		m.scanZoneIDs,
 		m.scanSort,
+		m.initRegions,
 		m.scanGateways,
 		m.scanCephMonitors,
 	} {
@@ -24,7 +32,7 @@ func (m *Cluster) scan(opts *ScanOptions) error {
 	return nil
 }
 
-func (m *Cluster) scanDirectory(_ *ScanOptions) error {
+func (m *Cluster) scanDirectory(opts *ScanOptions) error {
 	// each directory is a zone
 	entries, err := fs.ReadDir(m.dir, ".")
 	if err != nil {
@@ -33,16 +41,14 @@ func (m *Cluster) scanDirectory(_ *ScanOptions) error {
 
 	for _, e := range entries {
 		if e.IsDir() {
-			z, err := m.newZone(e.Name())
+			ok, err := m.scanSubdirectory(opts, e.Name())
 			switch {
 			case err != nil:
 				return core.Wrap(err, e.Name())
-			case z.Machines.Len() == 0:
-				z.warn(nil).
-					WithField("zone", z.Name).
+			case !ok:
+				m.warn(nil).
+					WithField("zone", e.Name()).
 					Print("empty")
-			default:
-				m.Zones = append(m.Zones, z)
 			}
 		}
 	}
@@ -50,6 +56,27 @@ func (m *Cluster) scanDirectory(_ *ScanOptions) error {
 	return nil
 }
 
+func (m *Cluster) scanSubdirectory(_ *ScanOptions, name string) (bool, error) {
+	z, err := m.newZone(name)
+	switch {
+	case err != nil:
+		// somewhere went wrong scanning the subdirectory
+		return false, err
+	case z.Machines.Len() > 0:
+		// zones have machines and the regions they belong
+		m.Zones = append(m.Zones, z)
+		return true, nil
+	case len(z.Regions) > 0:
+		// regions have no machines but can include
+		// other regions
+		m.appendRegionRegions(name, z.Regions...)
+		return true, nil
+	default:
+		// empty
+		return false, nil
+	}
+}
+
 func (m *Cluster) newZone(name string) (*Zone, error) {
 	z := &Zone{
 		zones:  m,
@@ -73,6 +100,10 @@ func (m *Cluster) scanMachines(opts *ScanOptions) error {
 		err = p.scan(opts)
 		return err != nil
 	})
+	m.ForEachMachine(func(p *Machine) bool {
+		err = p.scanWrapUp(opts)
+		return err != nil
+	})
 	return err
 }
 
@@ -150,34 +181,68 @@ func (z *Zone) scan() error {
 	}
 
 	for _, e := range entries {
-		if e.IsDir() {
-			m := &Machine{
-				zone:   z,
-				logger: z,
-				Name:   e.Name(),
-			}
+		name := e.Name()
 
-			m.debug().
-				WithField("node", m.Name).
+		switch {
+		case name == ZoneRegionsFileName:
+			err = z.loadRegions()
+		case e.IsDir():
+			err = z.scanSubdirectory(name)
+		default:
+			z.warn(nil).
 				WithField("zone", z.Name).
-				Print("found")
+				WithField("filename", name).
+				Print("unknown")
+		}
 
-			if err := m.init(); err != nil {
-				m.error(err).
-					WithField("node", m.Name).
-					WithField("zone", z.Name).
-					Print()
-
-				return err
-			}
-
-			z.Machines = append(z.Machines, m)
+		if err != nil {
+			return err
 		}
 	}
 
 	return nil
 }
 
+func (z *Zone) loadRegions() error {
+	filename := path.Join(z.Name, ZoneRegionsFileName)
+	regions, err := z.zones.ReadLines(filename)
+
+	if err == nil {
+		// parsed
+		err = z.appendRegions(regions...)
+		if err != nil {
+			err = core.Wrap(err, filename)
+		}
+	}
+
+	return err
+}
+
+func (z *Zone) scanSubdirectory(name string) error {
+	m := &Machine{
+		zone:   z,
+		logger: z,
+		Name:   name,
+	}
+
+	m.debug().
+		WithField("node", m.Name).
+		WithField("zone", z.Name).
+		Print("found")
+
+	if err := m.init(); err != nil {
+		m.error(err).
+			WithField("node", m.Name).
+			WithField("zone", z.Name).
+			Print()
+
+		return err
+	}
+
+	z.Machines = append(z.Machines, m)
+	return nil
+}
+
 // GetGateway returns the first gateway found, if none
 // files will be created to enable the first [Machine] to
 // be one

pkg/cluster/cluster_scan_options.go

@@ -6,7 +6,6 @@ import (
 
 	"darvaza.org/resolver"
 	"darvaza.org/slog"
-	"github.com/hack-pad/hackpadfs/os"
 )
 
 // A ScanOption pre-configures the Zones before scanning
@@ -28,7 +27,7 @@ type ScanOptions struct {
 // the DNS resolver to get PublicAddresses of nodes.
 // Default is true
 func ResolvePublicAddresses(resolve bool) ScanOption {
-	return func(m *Cluster, opt *ScanOptions) error {
+	return func(_ *Cluster, opt *ScanOptions) error {
 		opt.DontResolvePublicAddresses = !resolve
 		return nil
 	}
@@ -37,7 +36,7 @@ func ResolvePublicAddresses(resolve bool) ScanOption {
 // WithLookuper specifies what resolver.Lookuper to use to
 // find public addresses
 func WithLookuper(h resolver.Lookuper) ScanOption {
-	return func(m *Cluster, opt *ScanOptions) error {
+	return func(m *Cluster, _ *ScanOptions) error {
 		if h == nil {
 			return fs.ErrInvalid
 		}
@@ -50,7 +49,7 @@ func WithLookuper(h resolver.Lookuper) ScanOption {
 // public addresses. if nil is passed, the [net.Resolver] will be used.
 // The default is using Cloudflare's 1.1.1.1.
 func WithResolver(h resolver.Resolver) ScanOption {
-	return func(m *Cluster, opt *ScanOptions) error {
+	return func(m *Cluster, _ *ScanOptions) error {
 		if h == nil {
 			h = resolver.SystemResolver(true)
 		}
@@ -101,7 +100,7 @@ func NewFromDirectory(dir, domain string, opts ...ScanOption) (*Cluster, error)
 		return nil, err
 	}
 
-	sub, err := os.NewFS().Sub(fullPath[1:])
+	sub, err := DirFS(dir)
 	if err != nil {
 		return nil, err
 	}

pkg/cluster/hosts.go

@@ -0,0 +1,128 @@
+package cluster
+
+import (
+	"bytes"
+	"fmt"
+	"strings"
+	"text/template"
+)
+
+type hostsFile struct {
+	Ring0 []hostsEntry
+	Ring1 []hostsEntry
+}
+
+type hostsEntry struct {
+	Addr  string
+	Names []string
+}
+
+var hostsTemplate = template.Must(template.New("hosts").Funcs(template.FuncMap{
+	"StringsJoin": strings.Join,
+}).Parse(`127.0.0.1 localhost
+
+# The following lines are desirable for IPv6 capable hosts
+::1     ip6-localhost ip6-loopback
+fe00::0 ip6-localnet
+ff00::0 ip6-mcastprefix
+ff02::1 ip6-allnodes
+ff02::2 ip6-allrouters
+ff02::3 ip6-allhosts
+
+{{range .Ring1 -}}
+{{.Addr}}  {{StringsJoin .Names " "}}
+{{end}}
+{{range .Ring0 -}}
+{{.Addr}}  {{StringsJoin .Names " "}}
+{{end -}}
+`))
+
+// WriteHosts rewrites all hosts files on the tree
+func (m *Cluster) WriteHosts() error {
+	var err error
+
+	m.ForEachZone(func(z *Zone) bool {
+		err = z.WriteHosts()
+		return err != nil
+	})
+
+	return err
+}
+
+// WriteHosts rewrites all hosts files in the zone
+func (z *Zone) WriteHosts() error {
+	var err error
+
+	s := z.Hosts()
+	z.ForEachMachine(func(p *Machine) bool {
+		err = p.WriteStringFile(s, "hosts")
+		return err != nil
+	})
+
+	return err
+}
+
+// WriteHosts rewrites the hosts file
+func (p *Machine) WriteHosts() error {
+	s := p.zone.Hosts()
+	return p.WriteStringFile(s, "hosts")
+}
+
+func (z *Zone) genHosts(out *hostsFile, p *Machine) {
+	var names []string
+
+	ip, _ := RingOneAddress(p.zone.ID, p.ID)
+	names = append(names, p.Name)
+
+	if p.CephMonitor {
+		names = append(names, fmt.Sprintf("%s-%s", p.zone.Name, "ceph"))
+		names = append(names, fmt.Sprintf("%s-%s", p.zone.Name, "k3s"))
+
+		if z.ID == p.zone.ID {
+			names = append(names, "ceph")
+			names = append(names, "k3s")
+		}
+	}
+
+	entry := hostsEntry{
+		Addr:  ip.String(),
+		Names: names,
+	}
+
+	out.Ring1 = append(out.Ring1, entry)
+
+	if p.IsGateway() {
+		var s string
+
+		ip, _ = RingZeroAddress(p.zone.ID, p.ID)
+		s = fmt.Sprintf("%s-%v", p.Name, 0)
+
+		entry = hostsEntry{
+			Addr:  ip.String(),
+			Names: []string{s},
+		}
+
+		out.Ring0 = append(out.Ring0, entry)
+	}
+}
+
+// Hosts renders the /etc/hosts to be used on this zone
+func (z *Zone) Hosts() string {
+	var buf bytes.Buffer
+	var out hostsFile
+
+	z.zones.ForEachZone(func(z2 *Zone) bool {
+		z2.ForEachMachine(func(p *Machine) bool {
+			z.genHosts(&out, p)
+
+			return false
+		})
+		return false
+	})
+
+	if err := hostsTemplate.Execute(&buf, &out); err != nil {
+		panic(err)
+	}
+
+	return buf.String()
+}

pkg/cluster/machine.go

@@ -15,6 +15,7 @@ type Machine struct {
 	ID   int
 	Name string `json:"-" yaml:"-"`
 
+	Inactive        bool         `json:"inactive,omitempty"     yaml:"inactive,omitempty"`
 	CephMonitor     bool         `json:"ceph_monitor,omitempty" yaml:"ceph_monitor,omitempty"`
 	PublicAddresses []netip.Addr `json:"public,omitempty"       yaml:"public,omitempty"`
 	Rings           []*RingInfo  `json:"rings,omitempty"        yaml:"rings,omitempty"`
@@ -43,6 +44,11 @@ func (m *Machine) FullName() string {
 	return strings.Join(name, ".")
 }
 
+// IsActive indicates the machine is to be included in regions' DNS entries
+func (m *Machine) IsActive() bool {
+	return !m.Inactive
+}
+
 // IsGateway tells if the Machine is a ring0 gateway
 func (m *Machine) IsGateway() bool {
 	_, ok := m.getRingInfo(0)

pkg/cluster/machine_file.go

@@ -1,7 +1,6 @@
 package cluster
 
 import (
-	"bytes"
 	"fmt"
 	"io"
 	"os"
@@ -12,10 +11,9 @@ import (
 
 // OpenFile opens a file on the machine's config directory with the specified flags
 func (m *Machine) OpenFile(name string, flags int, args ...any) (fs.File, error) {
-	base := m.zone.zones.dir
 	fullName := m.getFilename(name, args...)
 
-	return fs.OpenFile(base, fullName, flags, 0644)
+	return m.zone.zones.OpenFile(fullName, flags)
 }
 
 // CreateTruncFile creates or truncates a file on the machine's config directory
@@ -34,42 +32,47 @@ func (m *Machine) openWriter(name string, flags int, args ...any) (io.WriteClose
 		return nil, err
 	}
 
-	return f.(io.WriteCloser), nil
+	if f, ok := f.(io.WriteCloser); ok {
+		return f, nil
+	}
+
+	panic("unreachable")
 }
 
 // RemoveFile deletes a file from the machine's config directory
 func (m *Machine) RemoveFile(name string, args ...any) error {
-	base := m.zone.zones.dir
 	fullName := m.getFilename(name, args...)
-	err := fs.Remove(base, fullName)
 
-	switch {
-	case os.IsNotExist(err):
-		return nil
-	default:
-		return err
-	}
+	return m.zone.zones.RemoveFile(fullName)
 }
 
 // ReadFile reads a file from the machine's config directory
 func (m *Machine) ReadFile(name string, args ...any) ([]byte, error) {
-	base := m.zone.zones.dir
 	fullName := m.getFilename(name, args...)
 
-	return fs.ReadFile(base, fullName)
+	return m.zone.zones.ReadFile(fullName)
+}
+
+// ReadLines reads a file from the machine's config directory,
+// split by lines, trimmed, and accepting `#` to comment lines out.
+func (m *Machine) ReadLines(name string, args ...any) ([]string, error) {
+	fullName := m.getFilename(name, args...)
+
+	return m.zone.zones.ReadLines(fullName)
 }
 
 // WriteStringFile writes the given content to a file on the machine's config directory
 func (m *Machine) WriteStringFile(value string, name string, args ...any) error {
-	f, err := m.CreateTruncFile(name, args...)
-	if err != nil {
-		return err
-	}
-	defer f.Close()
+	fullName := m.getFilename(name, args...)
 
-	buf := bytes.NewBufferString(value)
-	_, err = buf.WriteTo(f)
-	return err
+	return m.zone.zones.WriteStringFile(value, fullName)
+}
+
+// MkdirAll creates directories relative to the machine's config directory
+func (m *Machine) MkdirAll(name string, args ...any) error {
+	fullName := m.getFilename(name, args...)
+
+	return m.zone.zones.MkdirAll(fullName)
 }
 
 func (m *Machine) getFilename(name string, args ...any) string {

pkg/cluster/machine_rings.go

@@ -28,7 +28,7 @@ func (m *Machine) GetWireguardKeys(ring int) (wireguard.KeyPair, error) {
 	out.PrivateKey, err = wireguard.PrivateKeyFromBase64(string(data))
 	if err != nil {
 		// bad key
-		err = core.Wrapf(err, "wg%v.key", ring)
+		err = core.Wrap(err, "wg%v.key", ring)
 		return out, err
 	}
 
@@ -44,7 +44,7 @@ func (m *Machine) GetWireguardKeys(ring int) (wireguard.KeyPair, error) {
 		out.PublicKey, err = wireguard.PublicKeyFromBase64(string(data))
 		if err != nil {
 			// bad key
-			err = core.Wrapf(err, "wg%v.pub", ring)
+			err = core.Wrap(err, "wg%v.pub", ring)
 			return out, err
 		}
 	}
@@ -118,21 +118,31 @@ func (m *Machine) tryApplyWireguardConfig(ring int) error {
 	}
 }
 
-func (m *Machine) applyWireguardConfig(ring int, wg *wireguard.Config) error {
+func (m *Machine) applyWireguardConfigNode(ring int, wg *wireguard.Config) error {
 	addr := wg.GetAddress()
-	zoneID, nodeID, ok := Rings[ring].Decode(addr)
-	if !ok {
-		return fmt.Errorf("%s: invalid address", addr)
-	}
+	if !core.IsZero(addr) {
+		zoneID, nodeID, ok := Rings[ring].Decode(addr)
+		if !ok {
+			return fmt.Errorf("%s: invalid address", addr)
+		}
 
-	if err := m.applyZoneNodeID(zoneID, nodeID); err != nil {
-		return core.Wrapf(err, "%s: invalid address", addr)
+		if err := m.applyZoneNodeID(zoneID, nodeID); err != nil {
+			return core.Wrap(err, "%s: invalid address", addr)
+		}
 	}
 
 	if err := m.applyWireguardInterfaceConfig(ring, wg.Interface); err != nil {
 		return core.Wrap(err, "interface")
 	}
 
+	return nil
+}
+
+func (m *Machine) applyWireguardConfig(ring int, wg *wireguard.Config) error {
+	if err := m.applyWireguardConfigNode(ring, wg); err != nil {
+		return err
+	}
+
 	for _, peer := range wg.Peer {
 		err := m.applyWireguardPeerConfig(ring, peer)
 		switch {
@@ -230,6 +240,23 @@ func (m *Machine) applyZoneNodeID(zoneID, nodeID int) error {
 	return nil
 }
 
+func (m *Machine) setRingDefaults(ri *RingInfo) error {
+	if ri.Keys.PrivateKey.IsZero() {
+		m.info().
+			WithField("subsystem", "wireguard").
+			WithField("node", m.Name).
+			WithField("ring", ri.Ring).
+			Print("generating key pair")
+
+		kp, err := wireguard.NewKeyPair()
+		if err != nil {
+			return err
+		}
+		ri.Keys = kp
+	}
+	return nil
+}
+
 // RemoveWireguardConfig deletes wgN.conf from the machine's
 // config directory.
 func (m *Machine) RemoveWireguardConfig(ring int) error {

pkg/cluster/machine_scan.go

@@ -3,6 +3,7 @@ package cluster
 import (
 	"context"
 	"net/netip"
+	"os"
 	"strconv"
 	"strings"
 	"time"
@@ -68,7 +69,8 @@ func (m *Machine) setID() error {
 	return nil
 }
 
-func (m *Machine) scan(opts *ScanOptions) error {
+// scan is called once we know about all zones and machine names
+func (m *Machine) scan(_ *ScanOptions) error {
 	for i := 0; i < RingsCount; i++ {
 		if err := m.tryApplyWireguardConfig(i); err != nil {
 			m.error(err).
@@ -80,6 +82,45 @@ func (m *Machine) scan(opts *ScanOptions) error {
 		}
 	}
 
+	return m.loadInactive()
+}
+
+func (m *Machine) loadInactive() error {
+	data, err := m.ReadLines("region")
+	switch {
+	case os.IsNotExist(err):
+		// no file
+		return nil
+	case err != nil:
+		// read error
+		return err
+	default:
+		// look for "none"
+		for _, r := range data {
+			switch r {
+			case "none":
+				m.Inactive = true
+			default:
+				m.Inactive = false
+			}
+		}
+		return nil
+	}
+}
+
+// scanWrapUp is called once all machines have been scanned
+func (m *Machine) scanWrapUp(opts *ScanOptions) error {
+	for _, ri := range m.Rings {
+		if err := m.setRingDefaults(ri); err != nil {
+			m.error(err).
+				WithField("subsystem", "wireguard").
+				WithField("node", m.Name).
+				WithField("ring", ri.Ring).
+				Print()
+			return err
+		}
+	}
+
 	if !opts.DontResolvePublicAddresses {
 		return m.UpdatePublicAddresses()
 	}

pkg/cluster/regions.go

@@ -0,0 +1,231 @@
+package cluster
+
+import (
+	"bytes"
+	"path/filepath"
+)
+
+var (
+	_ MachineIterator = (*Region)(nil)
+	_ ZoneIterator    = (*Region)(nil)
+)
+
+// Region represents a group of zones geographically related
+type Region struct {
+	m     *Cluster
+	zones []*Zone
+
+	Name    string
+	Regions []string `json:",omitempty" yaml:",omitempty"`
+}
+
+// ForEachRegion calls a function for each Region of the cluster
+// until instructed to terminate the loop
+func (m *Cluster) ForEachRegion(fn func(r *Region) bool) {
+	for i := range m.Regions {
+		r := &m.Regions[i]
+		if fn(r) {
+			return
+		}
+	}
+}
+
+// ForEachMachine calls a function for each Machine in the region
+// until instructed to terminate the loop
+func (r *Region) ForEachMachine(fn func(*Machine) bool) {
+	r.ForEachZone(func(z *Zone) bool {
+		var term bool
+
+		z.ForEachMachine(func(p *Machine) bool {
+			if p.IsActive() {
+				term = fn(p)
+			}
+			return term
+		})
+
+		return term
+	})
+}
+
+// ForEachZone calls a function for each Zone in the region
+// until instructed to terminate the loop
+func (r *Region) ForEachZone(fn func(*Zone) bool) {
+	for _, p := range r.zones {
+		if fn(p) {
+			// terminate
+			return
+		}
+	}
+}
+
+func (m *Cluster) initRegions(_ *ScanOptions) error {
+	regions := make(map[string][]*Zone)
+
+	// first regions defined by zones
+	m.ForEachZone(func(z *Zone) bool {
+		SortRegions(z.Regions)
+		for _, region := range z.Regions {
+			regions[region] = append(regions[region], z)
+		}
+
+		return false
+	})
+
+	// bind first level regions and their zones
+	for name, zones := range regions {
+		m.setRegionZones(name, zones...)
+	}
+
+	// and combine zones to produce larger regions
+	for i := range m.Regions {
+		r := &m.Regions[i]
+		m.finishRegion(r)
+	}
+
+	m.sortRegions()
+	return nil
+}
+
+func (m *Cluster) setRegionZones(name string, zones ...*Zone) {
+	for i := range m.Regions {
+		r := &m.Regions[i]
+
+		if r.Name == name {
+			// found
+			r.m = m
+			r.zones = zones
+			return
+		}
+	}
+
+	// new
+	m.Regions = append(m.Regions, Region{
+		m:     m,
+		zones: zones,
+		Name:  name,
+	})
+}
+
+func (m *Cluster) appendRegionRegions(name string, subs ...string) {
+	for i := range m.Regions {
+		r := &m.Regions[i]
+
+		if name == r.Name {
+			// found
+			r.Regions = append(r.Regions, subs...)
+			return
+		}
+	}
+
+	// new
+	m.Regions = append(m.Regions, Region{
+		Name:    name,
+		Regions: subs,
+	})
+}
+
+func (z *Zone) appendRegions(regions ...string) error {
+	for _, s := range regions {
+		// TODO: validate
+		z.debug().
+			WithField("zone", z.Name).
+			WithField("region", s).
+			Print("attached")
+
+		z.Regions = append(z.Regions, s)
+	}
+
+	return nil
+}
+
+func (m *Cluster) finishRegion(r *Region) {
+	if r.m != nil {
+		// ready
+		return
+	}
+
+	r.m = m
+	sub := []string{}
+	for _, name := range r.Regions {
+		r2, ok := m.getRegion(name)
+		if !ok {
+			m.warn(nil).WithField("region", name).Print("unknown region")
+			continue
+		}
+
+		sub = append(sub, r2.Name)
+		r.zones = append(r.zones, r2.zones...)
+	}
+	r.Regions = sub
+}
+
+func (m *Cluster) getRegion(name string) (*Region, bool) {
+	for i := range m.Regions {
+		r := &m.Regions[i]
+
+		if name == r.Name {
+			m.finishRegion(r)
+			return r, true
+		}
+	}
+
+	return nil, false
+}
+
+// SyncRegions writes to the file system the regions this [Zone]
+// belongs to.
+func (z *Zone) SyncRegions() error {
+	err := z.syncZoneRegions()
+	if err == nil {
+		z.ForEachMachine(func(p *Machine) bool {
+			if p.IsActive() {
+				err = p.RemoveFile("region")
+			} else {
+				err = p.WriteStringFile("none\n", "region")
+			}
+			return err != nil
+		})
+	}
+
+	return err
+}
+
+func (z *Zone) syncZoneRegions() error {
+	name := filepath.Join(z.Name, "regions")
+
+	if len(z.Regions) > 0 {
+		var buf bytes.Buffer
+
+		for _, s := range z.Regions {
+			_, _ = buf.WriteString(s)
+			_, _ = buf.WriteRune('\n')
+		}
+
+		return z.zones.WriteStringFile(buf.String(), name)
+	}
+
+	return z.zones.RemoveFile(name)
+}
+
+// SyncRegions writes to the file system the regions covered
+// by this meta-region
+func (r *Region) SyncRegions() error {
+	name := filepath.Join(r.Name, "regions")
+
+	if len(r.Regions) > 0 {
+		var buf bytes.Buffer
+
+		for _, s := range r.Regions {
+			_, _ = buf.WriteString(s)
+			_, _ = buf.WriteRune('\n')
+		}
+
+		if err := r.m.MkdirAll(r.Name); err != nil {
+			return err
+		}
+
+		return r.m.WriteStringFile(buf.String(), name)
+	}
+
+	return r.m.RemoveFile(name)
+}

pkg/cluster/regions_utils.go

@@ -0,0 +1,41 @@
+package cluster
+
+import "sort"
+
+// SortRegions sorts regions. first by length those 3-character
+// or shorter, and then by length. It's mostly aimed at
+// supporting ISO-3166 order
+func SortRegions(regions []string) []string {
+	sort.Slice(regions, func(i, j int) bool {
+		r1, r2 := regions[i], regions[j]
+		return regionLess(r1, r2)
+	})
+
+	return regions
+}
+
+func regionLess(r1, r2 string) bool {
+	switch {
+	case len(r1) < 4:
+		switch {
+		case len(r1) < len(r2):
+			return true
+		case len(r1) > len(r2):
+			return false
+		default:
+			return r1 < r2
+		}
+	case len(r2) < 4:
+		return false
+	default:
+		return r1 < r2
+	}
+}
+
+func (m *Cluster) sortRegions() {
+	sort.Slice(m.Regions, func(i, j int) bool {
+		r1 := m.Regions[i].Name
+		r2 := m.Regions[j].Name
+		return regionLess(r1, r2)
+	})
+}

pkg/cluster/rings.go

@@ -41,7 +41,7 @@ func (ri *RingInfo) Merge(alter *RingInfo) error {
 		// can't disable via Merge
 		return fmt.Errorf("invalid %s: %v → %v", "enabled", ri.Enabled, alter.Enabled)
 	case !canMergeKeyPairs(ri.Keys, alter.Keys):
-		// incompatible keypairs
+		// incompatible key pairs
 		return fmt.Errorf("invalid %s: %s ≠ %s", "keys", ri.Keys, alter.Keys)
 	}
 

pkg/cluster/sync.go

@@ -3,8 +3,11 @@ package cluster
 // SyncAll updates all config files
 func (m *Cluster) SyncAll() error {
 	for _, fn := range []func() error{
+		m.SyncMkdirAll,
 		m.SyncAllWireguard,
 		m.SyncAllCeph,
+		m.SyncAllRegions,
+		m.WriteHosts,
 	} {
 		if err := fn(); err != nil {
 			return err
@@ -14,6 +17,20 @@ func (m *Cluster) SyncAll() error {
 	return nil
 }
 
+// SyncMkdirAll creates the directories needed to store files
+// required to represent the cluster.
+func (m *Cluster) SyncMkdirAll() error {
+	err := m.MkdirAll(".")
+	if err == nil {
+		m.ForEachMachine(func(p *Machine) bool {
+			err = p.MkdirAll(".")
+			return err != nil
+		})
+	}
+
+	return err
+}
+
 // SyncAllWireguard updates all wireguard config files
 func (m *Cluster) SyncAllWireguard() error {
 	var err error
@@ -42,3 +59,24 @@ func (m *Cluster) SyncAllCeph() error {
 
 	return m.WriteCephConfig(cfg)
 }
+
+// SyncAllRegions rewrites all region data
+func (m *Cluster) SyncAllRegions() error {
+	var err error
+
+	m.ForEachZone(func(z *Zone) bool {
+		err := z.SyncRegions()
+		return err != nil
+	})
+
+	if err != nil {
+		return err
+	}
+
+	m.ForEachRegion(func(r *Region) bool {
+		err = r.SyncRegions()
+		return err != nil
+	})
+
+	return err
+}

pkg/cluster/zones.go

@@ -19,8 +19,9 @@ type Zone struct {
 	zones  *Cluster
 	logger `json:"-" yaml:"-"`
 
-	ID   int
-	Name string
+	ID      int
+	Name    string
+	Regions []string `json:",omitempty" yaml:",omitempty"`
 
 	Machines
 }

pkg/dns/add.go

@@ -0,0 +1,69 @@
+package dns
+
+import (
+	"context"
+	"net/netip"
+	"os"
+	"time"
+
+	"darvaza.org/core"
+	"github.com/libdns/libdns"
+)
+
+// Add adds a machine to the DNS records
+func (mgr *Manager) Add(ctx context.Context, name string, addrs ...netip.Addr) error {
+	// TODO: validate name
+
+	cur, err := mgr.GetRecords(ctx, name)
+	if err != nil {
+		return core.Wrap(err, "GetRecords")
+	}
+
+	// merge []SyncAddr for name
+	s := mgr.asSyncRecordsMap(cur)[name+mgr.suffix]
+	for _, addr := range addrs {
+		s = AppendSyncAddr(s, addr)
+	}
+
+	return mgr.addSyncAddr(ctx, name, s)
+}
+
+func (mgr *Manager) addSyncAddr(ctx context.Context, name string, s []SyncAddr) error {
+	var recs []libdns.Record
+
+	for _, a := range s {
+		recs = append(recs, libdns.Record{
+			ID:    a.ID,
+			Name:  name + mgr.suffix,
+			Type:  core.IIf(a.Addr.Is6(), "AAAA", "A"),
+			TTL:   time.Second,
+			Value: a.Addr.String(),
+		})
+	}
+
+	SortRecords(recs)
+	err := writeRecords(recs, os.Stdout)
+	if err != nil {
+		return err
+	}
+
+	_, err = mgr.p.SetRecords(ctx, mgr.domain, recs)
+	return err
+}
+
+// AppendSyncAddr appends a [netip.Addr] to a [SyncAddr] slice
+// if the address is new.
+func AppendSyncAddr(s []SyncAddr, addr netip.Addr) []SyncAddr {
+	for _, se := range s {
+		if se.Addr.Compare(addr) == 0 {
+			// found
+			return s
+		}
+	}
+
+	s = append(s, SyncAddr{
+		Addr: addr,
+		TTL:  time.Second,
+	})
+	return s
+}

pkg/dns/dns.go

@@ -0,0 +1,38 @@
+// Package dns manages DNS entries for the cluster
+package dns
+
+import (
+	"fmt"
+	"net/netip"
+)
+
+// Zone represents a set of hosts with high affinity
+type Zone struct {
+	Name string
+
+	Hosts map[int]*Host
+}
+
+func (z *Zone) String() string {
+	if z == nil {
+		return "undetermined"
+	}
+	return z.Name
+}
+
+// Host represents a member of the cluster
+type Host struct {
+	zone *Zone
+
+	ID     int
+	Active bool
+	Addrs  []netip.Addr
+}
+
+func (p *Host) String() string {
+	if p == nil {
+		return "undetermined"
+	}
+
+	return fmt.Sprintf("%s-%v", p.zone, p.ID)
+}

pkg/dns/error.go

@@ -0,0 +1,12 @@
+package dns
+
+import "errors"
+
+var (
+	// ErrNoDNSProvider indicates a [libdns.Provider] wasn't assigned
+	// to the [Manager]
+	ErrNoDNSProvider = errors.New("dns provider not specified")
+
+	// ErrNoDomain indicates a domain wasn't specified
+	ErrNoDomain = errors.New("domain not specified")
+)

pkg/dns/manager.go

@@ -0,0 +1,230 @@
+package dns
+
+import (
+	"context"
+	"io/fs"
+	"net/netip"
+	"strings"
+
+	"darvaza.org/core"
+	"darvaza.org/slog"
+	"github.com/libdns/libdns"
+	"golang.org/x/net/publicsuffix"
+
+	"git.jpi.io/amery/jpictl/pkg/cluster"
+)
+
+// Manager is a DNS Manager instance
+type Manager struct {
+	domain  string
+	suffix  string
+	zones   map[string]*Zone
+	regions map[string][]string
+
+	p Provider
+	l slog.Logger
+}
+
+// ManagerOption configures a Manager
+type ManagerOption func(*Manager) error
+
+func newErrorManagerOption(err error, hint string) ManagerOption {
+	return func(*Manager) error {
+		return core.Wrap(err, hint)
+	}
+}
+
+// WithProvider attaches a libdns Provider to the Manager
+func WithProvider(p Provider) ManagerOption {
+	var err error
+
+	if p == nil {
+		p, err = DefaultDNSProvider()
+	}
+
+	if err != nil {
+		return newErrorManagerOption(err, "WithProvider")
+	}
+
+	return func(mgr *Manager) error {
+		mgr.p = p
+		return nil
+	}
+}
+
+// WithLogger attaches a logger to the Manager
+func WithLogger(log slog.Logger) ManagerOption {
+	if log == nil {
+		log = cluster.DefaultLogger()
+	}
+
+	return func(mgr *Manager) error {
+		mgr.l = log
+		return nil
+	}
+}
+
+func (mgr *Manager) setDefaults() error {
+	var opts []ManagerOption
+
+	if mgr.l == nil {
+		opts = append(opts, WithLogger(nil))
+	}
+
+	if mgr.domain == "" || mgr.suffix == "" {
+		return ErrNoDomain
+	}
+
+	for _, opt := range opts {
+		if err := opt(mgr); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// WithDomain specifies where the manager operates
+func WithDomain(domain string) ManagerOption {
+	base, err := publicsuffix.EffectiveTLDPlusOne(domain)
+	if err != nil {
+		return newErrorManagerOption(err, "publicsuffix")
+	}
+
+	suffix := strings.TrimSuffix(domain, base)
+	if suffix != "" {
+		suffix = "." + suffix[:len(suffix)-1]
+	}
+
+	return func(mgr *Manager) error {
+		mgr.domain = base
+		mgr.suffix = suffix
+		return nil
+	}
+}
+
+// NewManager creates a DNS manager with the
+func NewManager(opts ...ManagerOption) (*Manager, error) {
+	mgr := &Manager{
+		zones:   make(map[string]*Zone),
+		regions: make(map[string][]string),
+	}
+
+	for _, opt := range opts {
+		if err := opt(mgr); err != nil {
+			return nil, err
+		}
+	}
+
+	if err := mgr.setDefaults(); err != nil {
+		return nil, err
+	}
+	return mgr, nil
+}
+
+// GetRecords pulls all the address records on DNS for our domain,
+// optionally only those matching the given names.
+func (mgr *Manager) GetRecords(ctx context.Context, names ...string) ([]libdns.Record, error) {
+	if mgr.p == nil {
+		return nil, ErrNoDNSProvider
+	}
+
+	recs, err := mgr.p.GetRecords(ctx, mgr.domain)
+	switch {
+	case err != nil:
+		// failed
+		return nil, err
+	case len(recs) == 0:
+		// empty
+		return []libdns.Record{}, nil
+	case mgr.suffix == "" && len(names) == 0:
+		// unfiltered
+		return recs, nil
+	default:
+		// filtered
+		recs = mgr.filterRecords(recs, names...)
+		return recs, nil
+	}
+}
+
+func (mgr *Manager) filterRecords(recs []libdns.Record, names ...string) []libdns.Record {
+	out := make([]libdns.Record, 0, len(recs))
+	for _, rr := range recs {
+		name, ok := mgr.matchSuffix(rr)
+		switch {
+		case !ok:
+			// skip, wrong subdomain
+			continue
+		case len(names) == 0:
+			// unfiltered, take it
+		case !core.SliceContains(names, name):
+			// skip, not one of the requested names
+			continue
+		}
+
+		out = append(out, rr)
+	}
+
+	return out
+}
+
+func (mgr *Manager) matchSuffix(rr libdns.Record) (string, bool) {
+	if mgr.suffix == "" {
+		// no suffix
+		return rr.Name, true
+	}
+
+	// remove suffix
+	return strings.CutSuffix(rr.Name, mgr.suffix)
+}
+
+// AddHost registers a host
+func (mgr *Manager) AddHost(_ context.Context, zone string, id int,
+	active bool, addrs ...netip.Addr) error {
+	//
+	if zone == "" || id <= 0 {
+		return fs.ErrInvalid
+	}
+
+	z, ok := mgr.zones[zone]
+	if !ok {
+		z = &Zone{
+			Name:  zone,
+			Hosts: make(map[int]*Host),
+		}
+
+		mgr.zones[zone] = z
+	}
+
+	p := &Host{
+		zone:   z,
+		ID:     id,
+		Active: active,
+		Addrs:  SortAddrSlice(addrs),
+	}
+	z.Hosts[id] = p
+
+	if log, ok := mgr.l.Debug().WithEnabled(); ok {
+		log.WithField("subsystem", "dns").
+			WithField("zone", zone).
+			WithField("host", p.String()).
+			WithField("active", active).
+			Print()
+	}
+
+	return nil
+}
+
+// AddRegion specifies a new region and the zones it contains
+func (mgr *Manager) AddRegion(_ context.Context, region string, zones ...string) error {
+	mgr.regions[region] = append(mgr.regions[region], zones...)
+
+	if log, ok := mgr.l.Debug().WithEnabled(); ok {
+		for _, zoneName := range zones {
+			log.WithField("subsystem", "dns").
+				WithField("region", region).
+				WithField("zone", zoneName).Print()
+		}
+	}
+
+	return nil
+}

pkg/dns/provider.go

@@ -0,0 +1,38 @@
+package dns
+
+import (
+	"fmt"
+	"os"
+
+	"github.com/libdns/cloudflare"
+	"github.com/libdns/libdns"
+)
+
+const (
+	// CloudflareAPIToken is the environment variable
+	// containing the API Token
+	CloudflareAPIToken = "CLOUDFLARE_DNS_API_TOKEN"
+)
+
+// Provider manages DNS entries
+type Provider interface {
+	libdns.RecordGetter
+	libdns.RecordDeleter
+	libdns.RecordSetter
+	libdns.RecordAppender
+}
+
+// DefaultDNSProvider returns a cloudflare DNS provider
+// using an API Token from env [CloudflareAPIToken]
+func DefaultDNSProvider() (*cloudflare.Provider, error) {
+	s := os.Getenv(CloudflareAPIToken)
+	if s == "" {
+		return nil, fmt.Errorf("%q: %s", CloudflareAPIToken, "not found")
+	}
+
+	p := &cloudflare.Provider{
+		APIToken: s,
+	}
+
+	return p, nil
+}

pkg/dns/record.go

@@ -0,0 +1,247 @@
+package dns
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"net/netip"
+	"sort"
+	"strings"
+	"time"
+
+	"darvaza.org/core"
+	"github.com/libdns/libdns"
+
+	"git.jpi.io/amery/jpictl/pkg/cluster"
+)
+
+func (mgr *Manager) fqdn(name string) string {
+	return fmt.Sprintf("%s.%s.", name, mgr.domain)
+}
+
+// SortAddrSlice sorts a slice of [netip.Addr]
+func SortAddrSlice(s []netip.Addr) []netip.Addr {
+	sort.Slice(s, func(i, j int) bool {
+		return s[i].Less(s[j])
+	})
+	return s
+}
+
+// SortAddrRecords sorts a slice of [AddrRecord]
+// by Name and Address
+func SortAddrRecords(s []AddrRecord) []AddrRecord {
+	sort.Slice(s, func(i, j int) bool {
+		return s[i].Name < s[j].Name
+	})
+
+	for _, p := range s {
+		SortAddrSlice(p.Addr)
+	}
+
+	return s
+}
+
+// SortRecords sorts a slice of [libdns.Record], by Name, Type and Value
+func SortRecords(s []libdns.Record) []libdns.Record {
+	sort.Slice(s, func(i, j int) bool {
+		return lessRecord(s[i], s[j])
+	})
+	return s
+}
+
+func lessRecord(a, b libdns.Record) bool {
+	aName := strings.ToLower(a.Name)
+	bName := strings.ToLower(b.Name)
+
+	switch {
+	case aName < bName:
+		return true
+	case aName > bName:
+		return false
+	}
+
+	aType := strings.ToUpper(a.Type)
+	bType := strings.ToUpper(b.Type)
+
+	switch {
+	case aType < bType:
+		return true
+	case aType > bType:
+		return false
+	case aType == "A", aType == "AAAA":
+		// IP Addresses
+		var aa, ba netip.Addr
+
+		switch {
+		case aa.UnmarshalText([]byte(a.Value)) != nil:
+			// bad address on a
+			return true
+		case ba.UnmarshalText([]byte(b.Value)) != nil:
+			// bad address on b
+			return false
+		default:
+			return aa.Less(ba)
+		}
+	default:
+		// text
+		return a.Value < b.Value
+	}
+}
+
+// AddrRecord represents an A or AAAA record
+type AddrRecord struct {
+	Name string
+	Addr []netip.Addr
+}
+
+// Sort sorts the addresses of the record
+func (rr *AddrRecord) Sort() {
+	SortAddrSlice(rr.Addr)
+}
+
+// Export converts the record into libdns.Record
+func (rr *AddrRecord) Export() []libdns.Record {
+	out := make([]libdns.Record, len(rr.Addr))
+	for i, addr := range rr.Addr {
+		out[i] = libdns.Record{
+			Name:  rr.Name,
+			TTL:   time.Second * 1,
+			Type:  core.IIf(addr.Is6(), "AAAA", "A"),
+			Value: addr.String(),
+		}
+	}
+
+	return out
+}
+
+// WriteTo writes the record in BIND notation
+func (rr *AddrRecord) WriteTo(w io.Writer) (int64, error) {
+	var total int
+	for _, addr := range rr.Addr {
+		n, err := fmt.Fprint(w,
+			rr.Name, "\t",
+			1, "\t",
+			core.IIf(addr.Is6(), "AAAA", "A"), "\t",
+			addr.String(), "\n")
+
+		switch {
+		case err != nil:
+			return 0, err
+		case n > 0:
+			total += n
+		}
+	}
+
+	return int64(total), nil
+}
+
+// String converts the record into BIND entries
+func (rr *AddrRecord) String() string {
+	var buf bytes.Buffer
+	_, _ = rr.WriteTo(&buf)
+	return buf.String()
+}
+
+func (mgr *Manager) genRegionsSorted() []string {
+	regions := make([]string, 0, len(mgr.regions))
+	for name := range mgr.regions {
+		regions = append(regions, name)
+	}
+
+	return cluster.SortRegions(regions)
+}
+
+func (mgr *Manager) genZonesSorted() []string {
+	zones := make([]string, 0, len(mgr.zones))
+	for name := range mgr.zones {
+		zones = append(zones, name)
+	}
+
+	sort.Strings(zones)
+	return zones
+}
+
+func (mgr *Manager) genAllAddrRecords() []AddrRecord {
+	var out []AddrRecord
+
+	cache := make(map[string][]netip.Addr)
+
+	// zones
+	for _, z := range mgr.zones {
+		// hosts
+		s := mgr.genZoneHostRecords(z)
+		out = append(out, s...)
+
+		// zone alias
+		addrs := mgr.genZoneAddresses(z)
+		name := z.Name
+
+		out = append(out, AddrRecord{
+			Name: name + mgr.suffix,
+			Addr: addrs,
+		})
+
+		// and cache for regions
+		cache[name] = addrs
+	}
+
+	for _, name := range mgr.genRegionsSorted() {
+		var addrs []netip.Addr
+
+		for _, z := range mgr.regions[name] {
+			addrs = append(addrs, cache[z]...)
+		}
+
+		rec := AddrRecord{
+			Name: name + mgr.suffix,
+			Addr: addrs,
+		}
+		rec.Sort()
+
+		out = append(out, rec)
+	}
+
+	SortAddrRecords(out)
+	return out
+}
+
+func (*Manager) genZoneAddresses(z *Zone) []netip.Addr {
+	var out []netip.Addr
+
+	for _, p := range z.Hosts {
+		if p.Active {
+			out = append(out, p.Addrs...)
+		}
+	}
+
+	SortAddrSlice(out)
+	return out
+}
+
+func (mgr *Manager) genZoneHostRecords(z *Zone) []AddrRecord {
+	out := make([]AddrRecord, 0, len(z.Hosts))
+
+	for _, p := range z.Hosts {
+		rec := AddrRecord{
+			Name: p.String() + mgr.suffix,
+			Addr: p.Addrs,
+		}
+		out = append(out, rec)
+	}
+
+	SortAddrRecords(out)
+	return out
+}
+
+func (mgr *Manager) genRegionAddressesCached(name string,
+	zones map[string][]netip.Addr) []netip.Addr {
+	//
+	var addrs []netip.Addr
+
+	for _, zoneName := range mgr.regions[name] {
+		addrs = append(addrs, zones[zoneName]...)
+	}
+
+	SortAddrSlice(addrs)
+	return addrs
+}

pkg/dns/show.go

@@ -0,0 +1,58 @@
+package dns
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"io"
+	"os"
+	"time"
+
+	"darvaza.org/core"
+	"github.com/libdns/libdns"
+)
+
+// Show shows current DNS entries
+func (mgr *Manager) Show(ctx context.Context, names ...string) error {
+	recs, err := mgr.GetRecords(ctx, names...)
+	if err != nil {
+		return core.Wrap(err, "GetRecords")
+	}
+
+	SortRecords(recs)
+	return writeRecords(recs, os.Stdout)
+}
+
+func writeRecords(recs []libdns.Record, w io.Writer) error {
+	var buf bytes.Buffer
+
+	for _, rr := range recs {
+		_ = fmtRecord(&buf, rr)
+		_, _ = buf.WriteRune('\n')
+	}
+	_, _ = fmt.Fprintf(&buf, "; %v records\n", len(recs))
+
+	_, err := buf.WriteTo(w)
+	return err
+}
+
+func fmtRecord(w io.Writer, rr libdns.Record) error {
+	ttl := int(rr.TTL / time.Second)
+	if ttl < 1 {
+		ttl = 1
+	}
+
+	_, err := fmt.Fprintf(w, "%s\t%v\tIN\t%s\t%s",
+		rr.Name,
+		ttl,
+		rr.Type,
+		rr.Value)
+
+	if err == nil {
+		if rr.ID != "" {
+			_, err = fmt.Fprintf(w, "\t; %s", rr.ID)
+		}
+	}
+
+	return err
+}

pkg/dns/sync.go

@@ -0,0 +1,347 @@
+package dns
+
+import (
+	"context"
+	"net/netip"
+	"sort"
+	"strings"
+	"time"
+
+	"darvaza.org/core"
+	"darvaza.org/slog"
+	"github.com/libdns/libdns"
+)
+
+// SyncAddrRecord is similar to AddrRecord but include libdns.Record details
+// fetched from the Provider
+type SyncAddrRecord struct {
+	Name  string
+	Addrs []SyncAddr
+}
+
+// SyncAddr extends netip.Addr with ID and TTL fetched from the Provider
+type SyncAddr struct {
+	ID   string
+	Addr netip.Addr
+	TTL  time.Duration
+}
+
+// Export assembles a libdns.Record
+func (rec *SyncAddr) Export(name string) libdns.Record {
+	return libdns.Record{
+		ID:    rec.ID,
+		Name:  name,
+		Type:  core.IIf(rec.Addr.Is6(), "AAAA", "A"),
+		TTL:   time.Second,
+		Value: rec.Addr.String(),
+	}
+}
+
+// SortSyncAddrSlice sorts a slice of [SyncAddr] by its address
+func SortSyncAddrSlice(s []SyncAddr) []SyncAddr {
+	sort.Slice(s, func(i, j int) bool {
+		a1 := s[i].Addr
+		a2 := s[j].Addr
+		return a1.Less(a2)
+	})
+	return s
+}
+
+// GetSyncRecords pulls all the address records on DNS for our domain
+func (mgr *Manager) GetSyncRecords(ctx context.Context) ([]SyncAddrRecord, error) {
+	recs, err := mgr.GetRecords(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	return mgr.asSyncRecords(recs)
+}
+
+// AsSyncAddr converts a A or AAAA [libdns.Record] into a [SyncAddr]
+func (mgr *Manager) AsSyncAddr(rr libdns.Record) (SyncAddr, bool, error) {
+	var out SyncAddr
+	var addr netip.Addr
+
+	// skip non-address types
+	if rr.Type != "A" && rr.Type != "AAAA" {
+		return out, false, nil
+	}
+
+	// skip entries not containing our suffix
+	if mgr.suffix != "" {
+		if !strings.HasSuffix(rr.Name, mgr.suffix) {
+			return out, false, nil
+		}
+	}
+
+	err := addr.UnmarshalText([]byte(rr.Value))
+	if err != nil {
+		// invalid address on A or AAAA record
+		return out, false, err
+	}
+
+	out = SyncAddr{
+		ID:   rr.ID,
+		TTL:  rr.TTL,
+		Addr: addr,
+	}
+
+	return out, true, nil
+}
+
+func (mgr *Manager) asSyncRecordsMap(recs []libdns.Record) map[string][]SyncAddr {
+	// filter and convert
+	out := make(map[string][]SyncAddr)
+	for _, rr := range recs {
+		addr, ok, err := mgr.AsSyncAddr(rr)
+		switch {
+		case err != nil:
+			// skip invalid addresses
+			mgr.l.Error().
+				WithField("subsystem", "dns").
+				WithField(slog.ErrorFieldName, err).
+				WithField("name", rr.Name).
+				WithField("type", rr.Type).
+				WithField("addr", rr.Value).
+				Print()
+		case ok:
+			// store
+			out[rr.Name] = append(out[rr.Name], addr)
+		}
+	}
+	return out
+}
+
+func (mgr *Manager) asSyncRecords(recs []libdns.Record) ([]SyncAddrRecord, error) {
+	cache := mgr.asSyncRecordsMap(recs)
+
+	// prepare records
+	out := make([]SyncAddrRecord, len(cache))
+	names := make([]string, 0, len(cache))
+	for name := range cache {
+		names = append(names, name)
+	}
+	sort.Strings(names)
+
+	for i, name := range names {
+		addrs := cache[name]
+
+		out[i] = SyncAddrRecord{
+			Name:  name,
+			Addrs: SortSyncAddrSlice(addrs),
+		}
+	}
+
+	return out, nil
+}
+
+// Sync updates all the address records on DNS for our domain
+func (mgr *Manager) Sync(ctx context.Context) error {
+	current, err := mgr.GetSyncRecords(ctx)
+	if err != nil {
+		return core.Wrap(err, "GetRecords")
+	}
+
+	goal := mgr.genAllAddrRecords()
+	for _, p := range makeSyncMap(current, goal) {
+		err := mgr.doSync(ctx, p.Name, p.Before, p.After)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (mgr *Manager) doSync(ctx context.Context, name string,
+	before []SyncAddr, after []netip.Addr) error {
+	//
+	var err error
+
+	for _, a := range after {
+		before, err = mgr.doSyncUpdateOrInsert(ctx, name, a, before)
+		if err != nil {
+			return err
+		}
+	}
+
+	for _, b := range before {
+		err = mgr.doSyncRemove(ctx, name, b)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (mgr *Manager) doSyncUpdateOrInsert(ctx context.Context, name string,
+	addr netip.Addr, addrs []SyncAddr) ([]SyncAddr, error) {
+	//
+	var err error
+
+	i, ok := findSyncAddrSorted(addr, addrs)
+	if ok {
+		rec := addrs[i]
+
+		addrs = append(addrs[:i], addrs[i+1:]...)
+		err = mgr.doSyncUpdate(ctx, name, addr, rec)
+	} else {
+		err = mgr.doSyncInsert(ctx, name, addr)
+	}
+
+	return addrs, err
+}
+
+func (mgr *Manager) doSyncUpdate(ctx context.Context, name string,
+	addr netip.Addr, rec SyncAddr) error {
+	//
+	var log slog.Logger
+	var msg string
+	var err error
+
+	if rec.TTL != time.Second {
+		// amend TTL
+
+		// TODO: batch updates
+		_, err = mgr.p.SetRecords(ctx, mgr.domain, []libdns.Record{
+			rec.Export(name),
+		})
+
+		if err == nil {
+			log = mgr.l.Info()
+			msg = "Updated"
+		} else {
+			log = mgr.l.Error().
+				WithField(slog.ErrorFieldName, err)
+			msg = "Failed"
+		}
+	} else {
+		log = mgr.l.Info()
+		msg = "OK"
+	}
+
+	log.
+		WithField("subsystem", "dns").
+		WithField("name", name).
+		WithField("addr", addr).
+		Print(msg)
+
+	return err
+}
+
+func (mgr *Manager) doSyncInsert(ctx context.Context, name string,
+	addr netip.Addr) error {
+	//
+	var log slog.Logger
+	var msg string
+
+	rec := libdns.Record{
+		Name:  name,
+		Type:  core.IIf(addr.Is6(), "AAAA", "A"),
+		TTL:   time.Second,
+		Value: addr.String(),
+	}
+
+	_, err := mgr.p.AppendRecords(ctx, mgr.domain, []libdns.Record{
+		rec,
+	})
+
+	if err != nil {
+		log = mgr.l.Error().
+			WithField(slog.ErrorFieldName, err)
+		msg = "Failed to Add"
+	} else {
+		log = mgr.l.Info()
+		msg = "Added"
+	}
+
+	log.
+		WithField("subsystem", "dns").
+		WithField("name", name).
+		WithField("addr", addr).
+		Print(msg)
+	return err
+}
+
+func (mgr *Manager) doSyncRemove(ctx context.Context, name string,
+	rec SyncAddr) error {
+	//
+	var log slog.Logger
+	var msg string
+
+	// TODO: batch deletes
+	_, err := mgr.p.DeleteRecords(ctx, mgr.domain, []libdns.Record{
+		rec.Export(name),
+	})
+
+	if err != nil {
+		log = mgr.l.Error().
+			WithField(slog.ErrorFieldName, err)
+		msg = "Failed to Delete"
+	} else {
+		log = mgr.l.Warn()
+		msg = "Deleted"
+	}
+
+	log.
+		WithField("subsystem", "dns").
+		WithField("name", name).
+		WithField("addr", rec.Addr).
+		Print(msg)
+	return err
+}
+
+func findSyncAddrSorted(target netip.Addr, addrs []SyncAddr) (int, bool) {
+	for i, a := range addrs {
+		switch target.Compare(a.Addr) {
+		case 0:
+			// match
+			return i, true
+		case -1:
+			// miss
+			return -1, false
+		default:
+			// next
+		}
+	}
+
+	return -1, false
+}
+
+type syncMapEntry struct {
+	Name   string
+	Before []SyncAddr
+	After  []netip.Addr
+}
+
+func makeSyncMap(current []SyncAddrRecord,
+	goal []AddrRecord) map[string]syncMapEntry {
+	//
+	data := make(map[string]syncMapEntry)
+
+	for _, cur := range current {
+		me, ok := data[cur.Name]
+		if !ok {
+			me = syncMapEntry{
+				Name: cur.Name,
+			}
+		}
+
+		me.Before = append(me.Before, cur.Addrs...)
+		data[cur.Name] = me
+	}
+
+	for _, rr := range goal {
+		me, ok := data[rr.Name]
+		if !ok {
+			me = syncMapEntry{
+				Name: rr.Name,
+			}
+		}
+		me.After = append(me.After, rr.Addr...)
+		data[rr.Name] = me
+	}
+
+	return data
+}

pkg/dns/write.go

@@ -0,0 +1,63 @@
+package dns
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"net/netip"
+)
+
+// WriteTo writes the DNS data for the cluster
+func (mgr *Manager) WriteTo(w io.Writer) (int64, error) {
+	var buf bytes.Buffer
+
+	cache := make(map[string][]netip.Addr)
+
+	// zones
+	for _, zoneName := range mgr.genZonesSorted() {
+		z := mgr.zones[zoneName]
+
+		mgr.writeZoneHosts(&buf, z)
+
+		// zone alias
+		addrs := mgr.genZoneAddresses(z)
+
+		rr := AddrRecord{
+			Name: mgr.fqdn(zoneName + mgr.suffix),
+			Addr: addrs,
+		}
+		_, _ = rr.WriteTo(&buf)
+
+		// and cache for regions
+		cache[zoneName] = addrs
+	}
+
+	// regions, sorted
+	for _, name := range mgr.genRegionsSorted() {
+		addrs := mgr.genRegionAddressesCached(name, cache)
+
+		mgr.writeRegionAddresses(&buf, name, addrs)
+	}
+
+	return buf.WriteTo(w)
+}
+
+func (mgr *Manager) writeZoneHosts(w io.Writer, z *Zone) {
+	_, _ = fmt.Fprintf(w, ";\n; %s\n;\n", z.Name)
+
+	for _, rr := range mgr.genZoneHostRecords(z) {
+		rr.Name = mgr.fqdn(rr.Name)
+		_, _ = rr.WriteTo(w)
+	}
+}
+
+func (mgr *Manager) writeRegionAddresses(w io.Writer, name string, addrs []netip.Addr) {
+	_, _ = fmt.Fprintf(w, "; %s\n", name)
+
+	rr := AddrRecord{
+		Name: mgr.fqdn(name + mgr.suffix),
+		Addr: addrs,
+	}
+
+	_, _ = rr.WriteTo(w)
+}

pkg/wireguard/config.go

@@ -107,6 +107,11 @@ func (ep EndpointAddress) String() string {
 	}
 }
 
+// UnmarshalText loads an endpoint address from text data
+func (ep *EndpointAddress) UnmarshalText(b []byte) error {
+	return ep.FromString(string(b))
+}
+
 // FromString sets the EndpointAddress from a given "[host]:port"
 func (ep *EndpointAddress) FromString(s string) error {
 	host, port, err := core.SplitHostPort(s)
@@ -151,7 +156,7 @@ func (v *intermediateConfig) Export() (*Config, error) {
 	for i := 0; i < peers; i++ {
 		p, err := v.ExportPeer(i)
 		if err != nil {
-			err = core.Wrapf(err, "Peer[%v]:", i)
+			err = core.Wrap(err, "Peer[%v]:", i)
 			return nil, err
 		}
 
@@ -175,10 +180,12 @@ func (p interfaceConfig) Export() (InterfaceConfig, error) {
 		ListenPort: p.ListenPort,
 	}
 
-	out.PrivateKey, err = PrivateKeyFromBase64(p.PrivateKey)
-	if err != nil {
-		err = core.Wrap(err, "PrivateKey")
-		return InterfaceConfig{}, err
+	if p.PrivateKey != "" {
+		out.PrivateKey, err = PrivateKeyFromBase64(p.PrivateKey)
+		if err != nil {
+			err = core.Wrap(err, "PrivateKey")
+			return InterfaceConfig{}, err
+		}
 	}
 
 	return out, nil

pkg/wireguard/keys.go

@@ -51,6 +51,28 @@ func (pub PublicKey) String() string {
 	}
 }
 
+// UnmarshalText loads the value from base64
+func (key *PrivateKey) UnmarshalText(b []byte) error {
+	v, err := PrivateKeyFromBase64(string(b))
+	if err != nil {
+		return err
+	}
+
+	*key = v
+	return nil
+}
+
+// UnmarshalText loads the value from base64
+func (pub *PublicKey) UnmarshalText(b []byte) error {
+	v, err := PublicKeyFromBase64(string(b))
+	if err != nil {
+		return err
+	}
+
+	*pub = v
+	return nil
+}
+
 // MarshalJSON encodes the key for JSON, omitting empty.
 func (key PrivateKey) MarshalJSON() ([]byte, error) {
 	return encodeKeyJSON(key.String())
@@ -183,20 +205,14 @@ type KeyPair struct {
 // Validate checks the PublicKey matches the PrivateKey,
 // and sets the PublicKey if missing
 func (kp *KeyPair) Validate() error {
-	keyLen := len(kp.PrivateKey)
-	pubLen := len(kp.PublicKey)
-
 	switch {
-	case keyLen != PrivateKeySize:
-		// bad private key
+	case kp.PrivateKey.IsZero():
+		// no private key
 		return ErrInvalidPrivateKey
-	case pubLen == 0:
+	case kp.PublicKey.IsZero():
 		// no public key, set it
 		kp.PublicKey = kp.PrivateKey.Public()
 		return nil
-	case pubLen != PublicKeySize:
-		// bad public key
-		return ErrInvalidPublicKey
 	case !kp.PrivateKey.Public().Equal(kp.PublicKey):
 		// wrong public key
 		return ErrInvalidPublicKey