Merge pull request 'zones: fix jpictl dump by explicitly omitting Machine.logger and Zone.logger' (#15)

Reviewed-on: #15

.vscode/settings.json

@@ -0,0 +1,9 @@
+{
+    "cSpell.words": [
+        "ceph",
+        "darvaza",
+        "gofrs",
+        "jpictl",
+        "zerolog"
+    ]
+}

cmd/jpictl/config.go

@@ -17,5 +17,6 @@ var cfg = &Config{
 func (cfg *Config) LoadZones(resolve bool) (*zones.Zones, error) {
 	return zones.New(cfg.Base, cfg.Domain,
 		zones.ResolvePublicAddresses(resolve),
+		zones.WithLogger(log),
 	)
 }

cmd/jpictl/env.go

@@ -16,7 +16,11 @@ var envCmd = &cobra.Command{
 			return err
 		}
 
-		_, err = m.Env(*envExport).WriteTo(os.Stdout)
+		env, err := m.Env(*envExport)
+		if err != nil {
+			return err
+		}
+		_, err = env.WriteTo(os.Stdout)
 		return err
 	},
 }

cmd/jpictl/log.go

@@ -7,6 +7,7 @@ import (
 	"darvaza.org/slog"
 )
 
+// TODO: make log level configurable via flags
 var (
 	log = zerolog.New(nil, slog.Debug)
 )

go.mod

@@ -3,11 +3,14 @@ module git.jpi.io/amery/jpictl
 go 1.19
 
 require (
-	darvaza.org/core v0.9.5
-	darvaza.org/resolver v0.5.2
-	darvaza.org/sidecar v0.0.0-20230721122716-b9c54b8adbaf
-	darvaza.org/slog v0.5.2
+	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
 	github.com/burntSushi/toml v0.3.1
+	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
@@ -17,8 +20,9 @@ require (
 )
 
 require (
-	darvaza.org/slog/handlers/filter v0.4.4 // indirect
-	darvaza.org/slog/handlers/zerolog v0.4.4 // indirect
+	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
 	github.com/BurntSushi/toml v1.3.2 // indirect
 	github.com/chavacava/garif v0.1.0 // indirect
 	github.com/fatih/color v1.15.0 // indirect
@@ -37,7 +41,7 @@ require (
 	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 v0.11.0 // 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
 	gopkg.in/warnings.v0 v0.1.2 // indirect

go.sum

@@ -1,15 +1,21 @@
-darvaza.org/core v0.9.5 h1:sS5pZFwicaxJIQixEiqkMr9GknVHYL+EbKDMkR/4jDM=
-darvaza.org/core v0.9.5/go.mod h1:O3tHBMlw+xB47uGh5CUx7dXAujBAMmD8BCRFPZmIw54=
-darvaza.org/resolver v0.5.2 h1:VjHhEr/MJBszeDb7tYlXQ9Bsyh4xrDR7Sd10WAmPD6k=
-darvaza.org/resolver v0.5.2/go.mod h1:fFvsVPEFeMzUIWlLG47Go/6uJYtRLb9R8HIgYg3uaxE=
-darvaza.org/sidecar v0.0.0-20230721122716-b9c54b8adbaf h1:ya5ZQicBb/GWll3rlqra8No7oJXks7y1m/cJGYBypv4=
-darvaza.org/sidecar v0.0.0-20230721122716-b9c54b8adbaf/go.mod h1:by+bPsMa7Rxc/ZYG1qBunrtKocv/DkrPBmyFlmq/j2Q=
-darvaza.org/slog v0.5.2 h1:8TG1WyHjOyh2vW6t3pjzZVaWzpko5MIIpeI7LWqHFvs=
-darvaza.org/slog v0.5.2/go.mod h1:HAkEpxTA/mkiLNUXJo5qsCh8EVCtA3evje8GAaCDWHI=
-darvaza.org/slog/handlers/filter v0.4.4 h1:b2e2T9fQzMdJ0ia+f6b7kw9/T9GFwhFCKob/2tqhGGU=
-darvaza.org/slog/handlers/filter v0.4.4/go.mod h1:cQlJWuolB6guLug09sX/8Zrzct++M6SPCGvXR37E7Cc=
-darvaza.org/slog/handlers/zerolog v0.4.4 h1:OR1ASvH1fBCq3t85t4OU6oJPPuqMB1tsDoSpsh6HVJU=
-darvaza.org/slog/handlers/zerolog v0.4.4/go.mod h1:t60TeEbFcMLo74CkXC2S0rKlnwF4ixZyBR4fqIJV1GE=
+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=
 github.com/BurntSushi/toml v1.3.2 h1:o7IhLm0Msx3BaB+n3Ag7L8EVlByGnpq14C4YWiu/gL8=
 github.com/BurntSushi/toml v1.3.2/go.mod h1:CxXYINrC8qIiEnFrOxCa7Jy5BFHlXnUU2pbicEuybxQ=
 github.com/burntSushi/toml v0.3.1 h1:Hu1cOEC2qtKULZJCzym5tyA35bZr3HREuolgiAzMlhY=
@@ -26,6 +32,8 @@ github.com/fatih/color v1.15.0/go.mod h1:0h5ZqXfHYED7Bhv2ZJamyIOUej9KtShiJESRwBD
 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=
+github.com/gofrs/uuid/v5 v5.0.0 h1:p544++a97kEL+svbcFbCQVM9KFu0Yo25UoISXGNNH9M=
+github.com/gofrs/uuid/v5 v5.0.0/go.mod h1:CDOjlDMVAtN56jqyRUZh58JT31Tiw7/oQyEXZV+9bD8=
 github.com/hack-pad/hackpadfs v0.2.1 h1:FelFhIhv26gyjujoA/yeFO+6YGlqzmc9la/6iKMIxMw=
 github.com/hack-pad/hackpadfs v0.2.1/go.mod h1:khQBuCEwGXWakkmq8ZiFUvUZz84ZkJ2KNwKvChs4OrU=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
@@ -84,8 +92,8 @@ golang.org/x/sys v0.0.0-20210630005230-0f9fa26af87c/go.mod h1:oPkhp1MJrh7nUepCBc
 golang.org/x/sys v0.0.0-20210927094055-39ccf1dd6fa6/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 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 v0.11.0 h1:eG7RXZHdqOJ1i+0lgLgCpSXAp6M3LYlAo6osgSi0xOM=
-golang.org/x/sys v0.11.0/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=

pkg/ceph/ceph.go

@@ -0,0 +1,2 @@
+// Package ceph deals with ceph config
+package ceph

pkg/ceph/config.go

@@ -0,0 +1,67 @@
+package ceph
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"net/netip"
+	"strings"
+
+	"github.com/gofrs/uuid/v5"
+
+	"asciigoat.org/ini/basic"
+)
+
+// Config represents a ceph.conf file
+type Config struct {
+	Global GlobalConfig `ini:"global"`
+}
+
+// GlobalConfig represents the [global] section of a ceph.conf file
+type GlobalConfig struct {
+	FSID           uuid.UUID    `ini:"fsid"`
+	Monitors       []string     `ini:"mon_initial_members,comma"`
+	MonitorsAddr   []netip.Addr `ini:"mon_host,comma"`
+	ClusterNetwork netip.Prefix `ini:"cluster_network"`
+}
+
+// WriteTo writes a Wireguard [Config] onto the provided [io.Writer]
+func (cfg *Config) WriteTo(w io.Writer) (int64, error) {
+	var buf bytes.Buffer
+
+	writeGlobalToBuffer(&buf, &cfg.Global)
+	return buf.WriteTo(w)
+}
+
+func writeGlobalToBuffer(w *bytes.Buffer, c *GlobalConfig) {
+	_, _ = w.WriteString("[global]\n")
+	_, _ = fmt.Fprintf(w, "%s = %s\n", "fsid", c.FSID.String())
+	_, _ = 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())
+}
+
+func joinAddrs(addrs []netip.Addr, sep string) string {
+	s := make([]string, len(addrs))
+
+	for i, addr := range addrs {
+		s[i] = addr.String()
+	}
+
+	return strings.Join(s, sep)
+}
+
+// NewConfigFromReader parses the ceph.conf file
+func NewConfigFromReader(r io.Reader) (*Config, error) {
+	doc, err := basic.Decode(r)
+	if err != nil {
+		return nil, err
+	}
+
+	cfg, err := newConfigFromDocument(doc)
+	if err != nil {
+		return nil, err
+	}
+
+	return cfg, nil
+}

pkg/ceph/config_parser.go

@@ -0,0 +1,110 @@
+package ceph
+
+import (
+	"io/fs"
+	"net/netip"
+
+	"asciigoat.org/ini/basic"
+	"asciigoat.org/ini/parser"
+
+	"darvaza.org/core"
+)
+
+var sectionMap = map[string]func(*Config, *basic.Section) error{
+	"global": loadGlobalConfSection,
+}
+
+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 h(out, src)
+}
+
+func loadGlobalConfSection(out *Config, src *basic.Section) error {
+	var cfg GlobalConfig
+
+	for _, field := range src.Fields {
+		if err := loadGlobalConfField(&cfg, field); err != nil {
+			return core.Wrap(err, "global")
+		}
+	}
+
+	out.Global = cfg
+	return nil
+}
+
+// revive:disable:cyclomatic
+// revive:disable:cognitive-complexity
+
+func loadGlobalConfField(cfg *GlobalConfig, field basic.Field) error {
+	// revive:enable:cyclomatic
+	// revive:enable:cognitive-complexity
+
+	// TODO: refactor when asciigoat's ini parser learns to do reflection
+
+	switch field.Key {
+	case "fsid":
+		if !core.IsZero(cfg.FSID) {
+			return core.Wrapf(fs.ErrInvalid, "duplicate field %q", field.Key)
+		}
+
+		err := cfg.FSID.UnmarshalText([]byte(field.Value))
+		switch {
+		case err != nil:
+			return core.Wrap(err, field.Key)
+		default:
+			return nil
+		}
+	case "mon_host":
+		entries, _ := parser.SplitCommaArray(field.Value)
+		for _, s := range entries {
+			var addr netip.Addr
+
+			if err := addr.UnmarshalText([]byte(s)); err != nil {
+				return core.Wrap(err, field.Key)
+			}
+
+			cfg.MonitorsAddr = append(cfg.MonitorsAddr, addr)
+		}
+		return nil
+	case "mon_initial_members":
+		entries, _ := parser.SplitCommaArray(field.Value)
+		cfg.Monitors = append(cfg.Monitors, entries...)
+		return nil
+	case "cluster_network":
+		if !core.IsZero(cfg.ClusterNetwork) {
+			err := core.Wrap(fs.ErrInvalid, "fields before the first section")
+			return err
+		}
+
+		err := cfg.ClusterNetwork.UnmarshalText([]byte(field.Value))
+		switch {
+		case err != nil:
+			return core.Wrap(err, field.Key)
+		default:
+			return nil
+		}
+	}
+	return nil
+}
+
+func newConfigFromDocument(doc *basic.Document) (*Config, error) {
+	var out Config
+
+	if len(doc.Global) > 0 {
+		err := core.Wrap(fs.ErrInvalid, "fields before the first section")
+		return nil, err
+	}
+
+	for i := range doc.Sections {
+		src := &doc.Sections[i]
+		if err := loadConfSection(&out, src); err != nil {
+			return nil, err
+		}
+	}
+
+	return &out, nil
+}

pkg/zones/ceph.go

@@ -0,0 +1,122 @@
+package zones
+
+import (
+	"bytes"
+	"net/netip"
+	"sort"
+
+	"darvaza.org/core"
+	"github.com/gofrs/uuid/v5"
+
+	"git.jpi.io/amery/jpictl/pkg/ceph"
+)
+
+// GetCephFSID returns our Ceph's FSID
+func (m *Zones) GetCephFSID() (uuid.UUID, error) {
+	if core.IsZero(m.CephFSID) {
+		// generate one
+		v, err := uuid.NewV4()
+		if err != nil {
+			return uuid.Nil, err
+		}
+		m.CephFSID = v
+	}
+	return m.CephFSID, nil
+}
+
+// GetCephConfig reads the ceph.conf file
+func (m *Zones) GetCephConfig() (*ceph.Config, error) {
+	data, err := m.ReadFile("ceph.conf")
+	if err != nil {
+		return nil, err
+	}
+
+	r := bytes.NewReader(data)
+	return ceph.NewConfigFromReader(r)
+}
+
+// WriteCephConfig writes the ceph.conf file
+func (m *Zones) WriteCephConfig(cfg *ceph.Config) error {
+	f, err := m.CreateTruncFile("ceph.conf")
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	_, err = cfg.WriteTo(f)
+	return err
+}
+
+// GenCephConfig prepares a ceph.Config using the cluster information
+func (m *Zones) GenCephConfig() (*ceph.Config, error) {
+	fsid, err := m.GetCephFSID()
+	if err != nil {
+		return nil, err
+	}
+
+	cfg := &ceph.Config{
+		Global: ceph.GlobalConfig{
+			FSID: fsid,
+			ClusterNetwork: netip.PrefixFrom(
+				netip.AddrFrom4([4]byte{10, 0, 0, 0}),
+				8,
+			),
+		},
+	}
+
+	m.ForEachZone(func(z *Zone) bool {
+		for _, p := range z.GetCephMonitors() {
+			addr, _ := RingOneAddress(z.ID, p.ID)
+
+			cfg.Global.Monitors = append(cfg.Global.Monitors, p.Name)
+			cfg.Global.MonitorsAddr = append(cfg.Global.MonitorsAddr, addr)
+		}
+		return false
+	})
+
+	return cfg, nil
+}
+
+// GetCephMonitors returns the set of Ceph monitors on
+// the zone
+func (z *Zone) GetCephMonitors() Machines {
+	var mons Machines
+	var first, second *Machine
+
+	z.ForEachMachine(func(p *Machine) bool {
+		switch {
+		case p.CephMonitor:
+			// it is a monitor
+			mons = append(mons, p)
+		case len(mons) > 0:
+			// zone has a monitor
+		case first == nil && !p.IsGateway():
+			// first option for monitor
+			first = p
+		case second == nil:
+			// second option for monitor
+			second = p
+		}
+
+		return false
+	})
+
+	switch {
+	case len(mons) > 0:
+		// ready
+	case first != nil:
+		// make first option our monitor
+		first.CephMonitor = true
+		mons = append(mons, first)
+	case second != nil:
+		// make second option our monitor
+		second.CephMonitor = true
+		mons = append(mons, second)
+	default:
+		// zone without machines??
+		panic("unreachable")
+	}
+
+	sort.Sort(mons)
+	return mons
+}

pkg/zones/ceph_scan.go

@@ -0,0 +1,183 @@
+package zones
+
+import (
+	"net/netip"
+	"os"
+	"strings"
+
+	"darvaza.org/core"
+	"git.jpi.io/amery/jpictl/pkg/ceph"
+)
+
+// CephMissingMonitorError is an error that contains ceph
+// monitors present in ceph.conf but not found on the cluster
+type CephMissingMonitorError struct {
+	Names []string
+	Addrs []netip.Addr
+}
+
+func (err *CephMissingMonitorError) appendName(name string) {
+	err.Names = append(err.Names, name)
+}
+
+func (err *CephMissingMonitorError) appendAddr(addr netip.Addr) {
+	err.Addrs = append(err.Addrs, addr)
+}
+
+// OK tells if this instance actual shouldn't be treated as an error
+func (err CephMissingMonitorError) OK() bool {
+	switch {
+	case len(err.Names) > 0:
+		return false
+	case len(err.Addrs) > 0:
+		return false
+	default:
+		return true
+	}
+}
+
+func (err CephMissingMonitorError) Error() string {
+	if !err.OK() {
+		var buf strings.Builder
+
+		_, _ = buf.WriteString("missing:")
+		err.writeNames(&buf)
+		err.writeAddrs(&buf)
+
+		return buf.String()
+	}
+
+	// no error
+	return ""
+}
+
+func (err *CephMissingMonitorError) writeNames(w *strings.Builder) {
+	if len(err.Names) > 0 {
+		_, _ = w.WriteString(" mon_initial_members:")
+		for i, name := range err.Names {
+			if i != 0 {
+				_, _ = w.WriteRune(',')
+			}
+			_, _ = w.WriteString(name)
+		}
+	}
+}
+
+func (err *CephMissingMonitorError) writeAddrs(w *strings.Builder) {
+	if len(err.Addrs) > 0 {
+		_, _ = w.WriteString(" mon_host:")
+		for i, addr := range err.Addrs {
+			if i != 0 {
+				_, _ = w.WriteRune(',')
+			}
+			_, _ = w.WriteString(addr.String())
+		}
+	}
+}
+
+// AsError returns nil if the instance is actually OK
+func (err *CephMissingMonitorError) AsError() error {
+	if err == nil || err.OK() {
+		return nil
+	}
+
+	return err
+}
+
+type cephScanTODO struct {
+	names map[string]bool
+	addrs map[string]bool
+}
+
+func (todo *cephScanTODO) checkMachine(p *Machine) bool {
+	// on ceph all addresses are ring1
+	ring1, _ := RingOneAddress(p.Zone(), p.ID)
+	addr := ring1.String()
+
+	if _, found := todo.names[p.Name]; found {
+		// found on the TODO by name
+		todo.names[p.Name] = true
+		todo.addrs[addr] = true
+		return true
+	}
+
+	if _, found := todo.addrs[addr]; found {
+		// found on the TODO by address
+		todo.names[p.Name] = true
+		todo.addrs[addr] = true
+		return true
+	}
+
+	return false
+}
+
+func (todo *cephScanTODO) Missing() error {
+	var check CephMissingMonitorError
+
+	for name, found := range todo.names {
+		if !found {
+			check.appendName(name)
+		}
+	}
+
+	for addr, found := range todo.addrs {
+		if !found {
+			var a netip.Addr
+			// it wouldn't be on the map if it wasn't valid
+			_ = a.UnmarshalText([]byte(addr))
+
+			check.appendAddr(a)
+		}
+	}
+
+	return check.AsError()
+}
+
+func newCephScanTODO(cfg *ceph.Config) *cephScanTODO {
+	todo := &cephScanTODO{
+		names: make(map[string]bool),
+		addrs: make(map[string]bool),
+	}
+
+	for _, name := range cfg.Global.Monitors {
+		todo.names[name] = false
+	}
+
+	for _, addr := range cfg.Global.MonitorsAddr {
+		todo.addrs[addr.String()] = false
+	}
+
+	return todo
+}
+
+func (m *Zones) scanCephMonitors(_ *ScanOptions) error {
+	cfg, err := m.GetCephConfig()
+	switch {
+	case os.IsNotExist(err):
+		err = nil
+	case err != nil:
+		return err
+	}
+
+	if cfg != nil {
+		// store FSID
+		m.CephFSID = cfg.Global.FSID
+
+		// flag monitors based on config
+		todo := newCephScanTODO(cfg)
+		m.ForEachMachine(func(p *Machine) bool {
+			p.CephMonitor = todo.checkMachine(p)
+			return false
+		})
+		if err := todo.Missing(); err != nil {
+			return core.Wrap(err, "ceph")
+		}
+	}
+
+	// make sure every zone has one
+	m.ForEachZone(func(z *Zone) bool {
+		_ = z.GetCephMonitors()
+		return false
+	})
+	return nil
+}

pkg/zones/env.go

@@ -11,15 +11,24 @@ import (
 type Env struct {
 	ZoneIterator
 
-	export bool
+	cephFSID string
+	export   bool
 }
 
 // Env returns a shell environment factory
-func (m *Zones) Env(export bool) *Env {
-	return &Env{
+func (m *Zones) Env(export bool) (*Env, error) {
+	fsid, err := m.GetCephFSID()
+	if err != nil {
+		return nil, err
+	}
+
+	env := &Env{
 		ZoneIterator: m,
+		cephFSID:     fsid.String(),
 		export:       export,
 	}
+
+	return env, nil
 }
 
 // Zones returns the list of Zone IDs
@@ -38,7 +47,12 @@ func (m *Env) Zones() []int {
 func (m *Env) WriteTo(w io.Writer) (int64, error) {
 	var buf bytes.Buffer
 
+	if m.cephFSID != "" {
+		m.writeEnvVar(&buf, m.cephFSID, "FSID")
+	}
+
 	m.writeEnvVarInts(&buf, m.Zones(), "ZONES")
+
 	m.ForEachZone(func(z *Zone) bool {
 		m.writeEnvZone(&buf, z)
 		return false
@@ -59,6 +73,15 @@ func (m *Env) writeEnvZone(w io.Writer, z *Zone) {
 	// ZONE{zoneID}_GW
 	gateways, _ := z.GatewayIDs()
 	m.writeEnvVarInts(w, gateways, "ZONE%v_%s", zoneID, "GW")
+
+	// Ceph
+	monitors := z.GetCephMonitors()
+	// MON{zoneID}_NAME
+	m.writeEnvVar(w, genEnvZoneCephMonNames(monitors), "MON%v_%s", zoneID, "NAME")
+	// MON{zoneID}_IP
+	m.writeEnvVar(w, genEnvZoneCephMonIPs(monitors), "MON%v_%s", zoneID, "IP")
+	// MON{zoneID}_ID
+	m.writeEnvVar(w, genEnvZoneCephMonIDs(monitors), "MON%v_%s", zoneID, "ID")
 }
 
 func (m *Env) writeEnvVarInts(w io.Writer, value []int, name string, args ...any) {
@@ -111,3 +134,44 @@ func genEnvZoneNodes(z *Zone) string {
 	}
 	return ""
 }
+
+func genEnvZoneCephMonNames(m Machines) string {
+	var buf strings.Builder
+	m.ForEachMachine(func(p *Machine) bool {
+		if buf.Len() > 0 {
+			_, _ = buf.WriteRune(' ')
+		}
+		_, _ = buf.WriteString(p.Name)
+
+		return false
+	})
+	return buf.String()
+}
+
+func genEnvZoneCephMonIPs(m Machines) string {
+	var buf strings.Builder
+	m.ForEachMachine(func(p *Machine) bool {
+		addr, _ := RingOneAddress(p.Zone(), p.ID)
+
+		if buf.Len() > 0 {
+			_, _ = buf.WriteRune(' ')
+		}
+		_, _ = buf.WriteString(addr.String())
+
+		return false
+	})
+	return buf.String()
+}
+
+func genEnvZoneCephMonIDs(m Machines) string {
+	var buf strings.Builder
+	m.ForEachMachine(func(p *Machine) bool {
+		if buf.Len() > 0 {
+			_, _ = buf.WriteRune(' ')
+		}
+		_, _ = fmt.Fprintf(&buf, "%v", p.ID)
+
+		return false
+	})
+	return buf.String()
+}

pkg/zones/errors.go

@@ -0,0 +1,16 @@
+package zones
+
+import "errors"
+
+var (
+	// ErrInvalidName indicates the name isn't valid
+	ErrInvalidName = errors.New("invalid name")
+
+	// ErrUnknownNode indicates there is a reference to a node
+	// we don't have on the tree
+	ErrUnknownNode = errors.New("node does not exist")
+
+	// ErrInvalidNode indicates the nodes can't be used for
+	// the intended purpose
+	ErrInvalidNode = errors.New("invalid node")
+)

pkg/zones/log.go

@@ -0,0 +1,49 @@
+package zones
+
+import "darvaza.org/slog"
+
+type logger interface {
+	withDebug() (slog.Logger, bool)
+	withInfo() (slog.Logger, bool)
+
+	debug() slog.Logger
+	info() slog.Logger
+	warn(error) slog.Logger
+	error(error) slog.Logger
+}
+
+var (
+	_ logger = (*Zones)(nil)
+)
+
+func (z *Zones) withDebug() (slog.Logger, bool) {
+	return z.debug().WithEnabled()
+}
+
+func (z *Zones) withInfo() (slog.Logger, bool) {
+	return z.debug().WithEnabled()
+}
+
+func (z *Zones) debug() slog.Logger {
+	return z.log.Debug()
+}
+
+func (z *Zones) info() slog.Logger {
+	return z.log.Info()
+}
+
+func (z *Zones) warn(err error) slog.Logger {
+	l := z.log.Warn()
+	if err != nil {
+		l = l.WithField(slog.ErrorFieldName, err)
+	}
+	return l
+}
+
+func (z *Zones) error(err error) slog.Logger {
+	l := z.log.Error()
+	if err != nil {
+		l = l.WithField(slog.ErrorFieldName, err)
+	}
+	return l
+}

pkg/zones/machine.go

@@ -9,12 +9,16 @@ import (
 
 // A Machine is a machine on a Zone
 type Machine struct {
-	zone *Zone
+	zone   *Zone
+	logger `toml:"-" json:"-" yaml:"-"`
+
 	ID   int    `toml:"id"`
 	Name string `toml:"-" json:"-" yaml:"-"`
 
 	PublicAddresses []netip.Addr `toml:"public,omitempty" json:"public,omitempty" yaml:"public,omitempty"`
 	Rings           []*RingInfo  `toml:"rings,omitempty"  json:"rings,omitempty"  yaml:"rings,omitempty"`
+
+	CephMonitor bool `toml:"ceph_monitor,omitempty" json:"ceph_monitor,omitempty" yaml:"ceph_monitor,omitempty"`
 }
 
 // revive:enable:line-length-limit

pkg/zones/machine_rings.go

@@ -2,6 +2,7 @@ package zones
 
 import (
 	"bytes"
+	"errors"
 	"fmt"
 	"os"
 
@@ -135,7 +136,11 @@ func (m *Machine) applyWireguardConfig(ring int, wg *wireguard.Config) error {
 	}
 
 	for _, peer := range wg.Peer {
-		if err := m.applyWireguardPeerConfig(ring, peer); err != nil {
+		err := m.applyWireguardPeerConfig(ring, peer)
+		switch {
+		case errors.Is(err, ErrUnknownNode):
+			// ignore unknown peers
+		case err != nil:
 			err = core.Wrapf(err, "%s: wg%v:%s", m.Name, ring, addr)
 			return err
 		}
@@ -183,8 +188,10 @@ func (m *Machine) applyWireguardPeerConfig(ring int, pc wireguard.PeerConfig) er
 	switch {
 	case !found:
 		// unknown
+		return core.Wrap(ErrUnknownNode, pc.Endpoint.Host)
 	case ring == 1 && m.zone != peer.zone:
 		// invalid zone
+		return core.Wrap(ErrInvalidNode, peer.Name)
 	default:
 		// apply RingInfo
 		ri := &RingInfo{
@@ -197,8 +204,6 @@ func (m *Machine) applyWireguardPeerConfig(ring int, pc wireguard.PeerConfig) er
 
 		return peer.applyRingInfo(ring, ri)
 	}
-
-	return fmt.Errorf("%q: invalid peer endpoint", pc.Endpoint.Host)
 }
 
 func (m *Machine) applyZoneNodeID(zoneID, nodeID int) error {

pkg/zones/machine_scan.go

@@ -4,7 +4,10 @@ import (
 	"context"
 	"net/netip"
 	"strconv"
+	"strings"
 	"time"
+
+	"darvaza.org/core"
 )
 
 // LookupNetIP uses the DNS Resolver to get the public addresses associated
@@ -30,21 +33,32 @@ func (m *Machine) UpdatePublicAddresses() error {
 
 func (m *Machine) init() error {
 	if err := m.setID(); err != nil {
-		return err
+		return core.Wrap(err, m.Name)
 	}
 
 	for i := 0; i < RingsCount; i++ {
 		if err := m.tryReadWireguardKeys(i); err != nil {
-			return err
+			return core.Wrap(err, m.Name)
 		}
 	}
+
 	return nil
 }
 
 func (m *Machine) setID() error {
 	zoneName := m.zone.Name
-	suffix := m.Name[len(zoneName)+1:]
 
+	l := len(zoneName)
+	switch {
+	case len(m.Name) < l+2:
+		return ErrInvalidName
+	case !strings.HasPrefix(m.Name, zoneName):
+		return ErrInvalidName
+	case m.Name[l] != '-':
+		return ErrInvalidName
+	}
+
+	suffix := m.Name[l+1:]
 	id, err := strconv.ParseInt(suffix, 10, 8)
 	if err != nil {
 		return err

pkg/zones/options.go

@@ -5,10 +5,12 @@ import (
 	"path/filepath"
 
 	"darvaza.org/resolver"
+	"darvaza.org/slog"
+	"darvaza.org/slog/handlers/discard"
 	"github.com/hack-pad/hackpadfs/os"
 )
 
-// A ScanOption preconfigures the Zones before scanning
+// A ScanOption pre-configures the Zones before scanning
 type ScanOption func(*Zones, *ScanOptions) error
 
 // ScanOptions contains flags used by the initial scan
@@ -17,6 +19,10 @@ type ScanOptions struct {
 	// pre-populate Machine.PublicAddresses during the
 	// initial scan
 	DontResolvePublicAddresses bool
+
+	// Logger specifies the logger to be used. otherwise
+	// the scanner will be mute
+	slog.Logger
 }
 
 // ResolvePublicAddresses instructs the scanner to use
@@ -55,6 +61,19 @@ func WithResolver(h resolver.Resolver) ScanOption {
 	}
 }
 
+// WithLogger specifies what to use for logging
+func WithLogger(log slog.Logger) ScanOption {
+	return func(m *Zones, opt *ScanOptions) error {
+		if log == nil {
+			log = discard.New()
+		}
+
+		opt.Logger = log
+		m.log = log
+		return nil
+	}
+}
+
 func (m *Zones) setDefaults(opt *ScanOptions) error {
 	if m.resolver == nil {
 		h := resolver.NewCloudflareLookuper()
@@ -64,6 +83,12 @@ func (m *Zones) setDefaults(opt *ScanOptions) error {
 		}
 	}
 
+	if opt.Logger == nil {
+		if err := WithLogger(nil)(m, opt); err != nil {
+			return err
+		}
+	}
+
 	return nil
 }
 

pkg/zones/scan.go

@@ -3,6 +3,8 @@ package zones
 import (
 	"io/fs"
 	"sort"
+
+	"darvaza.org/core"
 )
 
 func (m *Zones) scan(opts *ScanOptions) error {
@@ -12,6 +14,7 @@ func (m *Zones) scan(opts *ScanOptions) error {
 		m.scanZoneIDs,
 		m.scanSort,
 		m.scanGateways,
+		m.scanCephMonitors,
 	} {
 		if err := fn(opts); err != nil {
 			return err
@@ -30,22 +33,32 @@ func (m *Zones) scanDirectory(_ *ScanOptions) error {
 
 	for _, e := range entries {
 		if e.IsDir() {
-			z := &Zone{
-				zones: m,
-				Name:  e.Name(),
+			z, err := m.newZone(e.Name())
+			switch {
+			case err != nil:
+				return core.Wrap(err, e.Name())
+			case z.Machines.Len() > 0:
+				m.Zones = append(m.Zones, z)
 			}
-
-			if err := z.scan(); err != nil {
-				return err
-			}
-
-			m.Zones = append(m.Zones, z)
 		}
 	}
 
 	return nil
 }
 
+func (m *Zones) newZone(name string) (*Zone, error) {
+	z := &Zone{
+		zones:  m,
+		logger: m,
+		Name:   name,
+	}
+
+	if err := z.scan(); err != nil {
+		return nil, err
+	}
+	return z, nil
+}
+
 func (m *Zones) scanMachines(opts *ScanOptions) error {
 	var err error
 	m.ForEachMachine(func(p *Machine) bool {
@@ -131,8 +144,9 @@ func (z *Zone) scan() error {
 	for _, e := range entries {
 		if e.IsDir() {
 			m := &Machine{
-				zone: z,
-				Name: e.Name(),
+				zone:   z,
+				logger: z,
+				Name:   e.Name(),
 			}
 
 			if err := m.init(); err != nil {

pkg/zones/sync.go

@@ -4,6 +4,7 @@ package zones
 func (m *Zones) SyncAll() error {
 	for _, fn := range []func() error{
 		m.SyncAllWireguard,
+		m.SyncAllCeph,
 	} {
 		if err := fn(); err != nil {
 			return err
@@ -31,3 +32,13 @@ func (m *Zones) SyncAllWireguard() error {
 
 	return nil
 }
+
+// SyncAllCeph updates the ceph.conf file
+func (m *Zones) SyncAllCeph() error {
+	cfg, err := m.GenCephConfig()
+	if err != nil {
+		return err
+	}
+
+	return m.WriteCephConfig(cfg)
+}

pkg/zones/zones.go

@@ -6,6 +6,8 @@ import (
 	"sort"
 
 	"darvaza.org/resolver"
+	"darvaza.org/slog"
+	"github.com/gofrs/uuid/v5"
 )
 
 var (
@@ -85,7 +87,8 @@ func FilterMachines(m MachineIterator, cond func(*Machine) bool) (Machines, int)
 
 // Zone represents one zone in a cluster
 type Zone struct {
-	zones *Zones
+	zones  *Zones
+	logger `toml:"-" json:"-" yaml:"-"`
 
 	ID   int    `toml:"id"`
 	Name string `toml:"name"`
@@ -135,15 +138,22 @@ func (z *Zone) GatewayIDs() ([]int, int) {
 	return out, len(out)
 }
 
+// revive:disable:line-length-limit
+
 // Zones represents all zones in a cluster
 type Zones struct {
 	dir      fs.FS
+	log      slog.Logger
 	resolver resolver.Resolver
 	domain   string
 
+	CephFSID uuid.UUID `toml:"ceph_fsid,omitempty" json:"ceph_fsid,omitempty" yaml:"ceph_fsid,omitempty"`
+
 	Zones []*Zone `toml:"zones"`
 }
 
+// revive:enable:line-length-limit
+
 // ForEachMachine calls a function for each Machine in the cluster
 // until instructed to terminate the loop
 func (m *Zones) ForEachMachine(fn func(*Machine) bool) {