merge master

Signed-off-by: husharp <[email protected]>

CHANGELOG.md

@@ -1,5 +1,25 @@
 TiUP Changelog
 
+## [1.14.1] 2024-01-12
+
+### Fixes
+
+- Fix tiproxy will restart even if version is the same in `tiup-cluster` (#2358, @xhebox)
+- Fix that empty label is not allowed in Prometheus in `tiup-cluster` (#2349, @borissavelev)
+- Fix can not set runtime config in config file & set TiFlash logger level to debug in `tiup-playground` (#2346, @Lloyd-Pottiger)
+- Fix TLS configs for tiproxy in `tiup-cluster` (#2356, @xhebox)
+- Fix cannot clone without global version in `tiup` (#2359, @nexustar)
+- Fix tiproxy config and version in `tiup-cluster` (#2354, @xhebox)
+- TiProxy needs to override usercfg in `tiup-playground` (#2342, @xhebox)
+- Fix the start script for alertmanager when enabling numa in `tiup-cluster` and `tiup-dm` (#2337, @KanShiori)
+- Fix metric targets will be overrided in `tiup-playground` (#2324, @xhebox)
+
+### Improvements
+
+- Change typo waitting to waiting in output messages for pdapi.go in `tiup-cluster` (#2330, @guangleibao)
+- Add check for rocky linux in `tiup-cluster` (#2333, @dveeden)
+- Fallback to default version when componen specific version is empty in `tiup-playground` (#2344, @xhebox)
+
 ## [1.14.0] 2023-11-17
 
 ### New Features

components/cluster/command/template.go

@@ -35,6 +35,7 @@ type TemplateOptions struct {
 type LocalTemplate struct {
 	GlobalUser          string   // global.user in yaml template
 	GlobalGroup         string   // global.group in yaml template
+	GlobalSystemdMode   string   // global.systemd_mode in yaml template
 	GlobalSSHPort       int      // global.ssh_port in yaml template
 	GlobalDeployDir     string   // global.deploy_dir in yaml template
 	GlobalDataDir       string   // global.data_dir in yaml template
@@ -122,6 +123,7 @@ func newTemplateCmd() *cobra.Command {
 	// template values for rendering
 	cmd.Flags().StringVar(&localOpt.GlobalUser, "user", "tidb", "The user who runs the tidb cluster.")
 	cmd.Flags().StringVar(&localOpt.GlobalGroup, "group", "", "group is used to specify the group name the user belong to if it's not the same as user.")
+	cmd.Flags().StringVar(&localOpt.GlobalSystemdMode, "systemd_mode", "system", "systemd_mode is used to select whether to use sudo permissions.")
 	cmd.Flags().IntVar(&localOpt.GlobalSSHPort, "ssh-port", 22, "SSH port of servers in the managed cluster.")
 	cmd.Flags().StringVar(&localOpt.GlobalDeployDir, "deploy-dir", "/tidb-deploy", "Storage directory for cluster deployment files, startup scripts, and configuration files.")
 	cmd.Flags().StringVar(&localOpt.GlobalDataDir, "data-dir", "/tidb-data", "TiDB Cluster data storage directory.")

components/dm/command/template.go

@@ -33,6 +33,7 @@ type TemplateOptions struct {
 type LocalTemplate struct {
 	GlobalUser          string   // global.user in yaml template
 	GlobalGroup         string   // global.group in yaml template
+	GlobalSystemdMode   string   // global.systemd_mode in yaml template
 	GlobalSSHPort       int      // global.ssh_port in yaml template
 	GlobalDeployDir     string   // global.deploy_dir in yaml template
 	GlobalDataDir       string   // global.data_dir in yaml template
@@ -113,6 +114,7 @@ func newTemplateCmd() *cobra.Command {
 	// template values for rendering
 	cmd.Flags().StringVar(&localOpt.GlobalUser, "user", "tidb", "The user who runs the tidb cluster.")
 	cmd.Flags().StringVar(&localOpt.GlobalGroup, "group", "", "group is used to specify the group name the user belong to if it's not the same as user.")
+	cmd.Flags().StringVar(&localOpt.GlobalSystemdMode, "systemd_mode", "system", "systemd_mode is used to select whether to use sudo permissions.")
 	cmd.Flags().IntVar(&localOpt.GlobalSSHPort, "ssh-port", 22, "SSH port of servers in the managed cluster.")
 	cmd.Flags().StringVar(&localOpt.GlobalDeployDir, "deploy-dir", "/tidb-deploy", "Storage directory for cluster deployment files, startup scripts, and configuration files.")
 	cmd.Flags().StringVar(&localOpt.GlobalDataDir, "data-dir", "/tidb-data", "TiDB Cluster data storage directory.")

components/playground/instance/tiflash.go

@@ -20,6 +20,7 @@ import (
 	"path/filepath"
 	"strings"
 
+	"github.com/pingcap/errors"
 	"github.com/pingcap/tiup/pkg/tidbver"
 	"github.com/pingcap/tiup/pkg/utils"
 )
@@ -103,9 +104,9 @@ func (inst *TiFlashInstance) StatusAddrs() (addrs []string) {
 }
 
 // Start calls set inst.cmd and Start
-func (inst *TiFlashInstance) Start(ctx context.Context) error {
-	if !tidbver.TiFlashPlaygroundNewStartMode(inst.Version.String()) {
-		return inst.startOld(ctx, inst.Version)
+func (inst *TiFlashInstance) Start(ctx context.Context, version utils.Version) error {
+	if !tidbver.TiFlashPlaygroundNewStartMode(version.String()) {
+		return inst.startOld(ctx, version)
 	}
 
 	proxyConfigPath := filepath.Join(inst.Dir, "tiflash_proxy.toml")
@@ -130,31 +131,59 @@ func (inst *TiFlashInstance) Start(ctx context.Context) error {
 
 	args := []string{
 		"server",
-	}
-	args = append(args,
 		fmt.Sprintf("--config-file=%s", configPath),
 		"--",
-		fmt.Sprintf("--tmp_path=%s", filepath.Join(inst.Dir, "tmp")),
-		fmt.Sprintf("--path=%s", filepath.Join(inst.Dir, "data")),
-		fmt.Sprintf("--listen_host=%s", inst.Host),
-		fmt.Sprintf("--logger.log=%s", inst.LogFile()),
-		fmt.Sprintf("--logger.errorlog=%s", filepath.Join(inst.Dir, "tiflash_error.log")),
-		fmt.Sprintf("--status.metrics_port=%d", inst.StatusPort),
-		fmt.Sprintf("--flash.service_addr=%s", utils.JoinHostPort(AdvertiseHost(inst.Host), inst.ServicePort)),
-		fmt.Sprintf("--raft.pd_addr=%s", strings.Join(endpoints, ",")),
-		fmt.Sprintf("--flash.proxy.addr=%s", utils.JoinHostPort(inst.Host, inst.ProxyPort)),
-		fmt.Sprintf("--flash.proxy.advertise-addr=%s", utils.JoinHostPort(AdvertiseHost(inst.Host), inst.ProxyPort)),
-		fmt.Sprintf("--flash.proxy.status-addr=%s", utils.JoinHostPort(inst.Host, inst.ProxyStatusPort)),
-		fmt.Sprintf("--flash.proxy.data-dir=%s", filepath.Join(inst.Dir, "proxy_data")),
-		fmt.Sprintf("--flash.proxy.log-file=%s", filepath.Join(inst.Dir, "tiflash_tikv.log")),
-	)
+	}
+	runtimeConfig := [][]string{
+		{"path", filepath.Join(inst.Dir, "data")},
+		{"listen_host", inst.Host},
+		{"logger.log", inst.LogFile()},
+		{"logger.errorlog", filepath.Join(inst.Dir, "tiflash_error.log")},
+		{"status.metrics_port", fmt.Sprintf("%d", inst.StatusPort)},
+		{"flash.service_addr", utils.JoinHostPort(AdvertiseHost(inst.Host), inst.ServicePort)},
+		{"raft.pd_addr", strings.Join(endpoints, ",")},
+		{"flash.proxy.addr", utils.JoinHostPort(inst.Host, inst.ProxyPort)},
+		{"flash.proxy.advertise-addr", utils.JoinHostPort(AdvertiseHost(inst.Host), inst.ProxyPort)},
+		{"flash.proxy.status-addr", utils.JoinHostPort(inst.Host, inst.ProxyStatusPort)},
+		{"flash.proxy.data-dir", filepath.Join(inst.Dir, "proxy_data")},
+		{"flash.proxy.log-file", filepath.Join(inst.Dir, "tiflash_tikv.log")},
+	}
+	userConfig, err := unmarshalConfig(configPath)
+	if err != nil {
+		return errors.Trace(err)
+	}
+	fmt.Println("userConfig", userConfig)
+	for _, arg := range runtimeConfig {
+		// if user has set the config, skip it
+		if !isKeyPresentInMap(userConfig, arg[0]) {
+			args = append(args, fmt.Sprintf("--%s=%s", arg[0], arg[1]))
+		}
+	}
 
 	inst.Process = &process{cmd: PrepareCommand(ctx, inst.BinPath, args, nil, inst.Dir)}
 
 	logIfErr(inst.Process.SetOutputFile(inst.LogFile()))
 	return inst.Process.Start()
 }
 
+func isKeyPresentInMap(m map[string]any, key string) bool {
+	keys := strings.Split(key, ".")
+	currentMap := m
+
+	for i := 0; i < len(keys); i++ {
+		if _, ok := currentMap[keys[i]]; !ok {
+			return false
+		}
+
+		// If the current value is a nested map, update the current map to the nested map
+		if innerMap, ok := currentMap[keys[i]].(map[string]any); ok {
+			currentMap = innerMap
+		}
+	}
+
+	return true
+}
+
 // Component return the component name.
 func (inst *TiFlashInstance) Component() string {
 	return "tiflash"

components/playground/instance/tiflash_config.go

@@ -28,6 +28,7 @@ func (inst *TiFlashInstance) getConfig() map[string]any {
 	config := make(map[string]any)
 
 	config["flash.proxy.config"] = filepath.Join(inst.Dir, "tiflash_proxy.toml")
+	config["logger.level"] = "debug"
 
 	if inst.Role == TiFlashRoleDisaggWrite {
 		config["storage.s3.endpoint"] = inst.DisaggOpts.S3Endpoint

components/playground/instance/tiproxy.go

@@ -65,7 +65,7 @@ func (c *TiProxy) MetricAddr() (r MetricAddr) {
 
 // Start implements Instance interface.
 func (c *TiProxy) Start(ctx context.Context) error {
-	endpoints := pdEndpoints(c.pds, true)
+	endpoints := pdEndpoints(c.pds, false)
 
 	configPath := filepath.Join(c.Dir, "config", "proxy.toml")
 	dir := filepath.Dir(configPath)
@@ -101,6 +101,10 @@ func (c *TiProxy) Start(ctx context.Context) error {
 		fmt.Sprintf("--config=%s", configPath),
 	}
 
+	if c.BinPath, err = tiupexec.PrepareBinary("tiproxy", version, c.BinPath); err != nil {
+		return err
+	}
+
 	c.Process = &process{cmd: PrepareCommand(ctx, c.BinPath, args, nil, c.Dir)}
 
 	logIfErr(c.Process.SetOutputFile(c.LogFile()))

components/playground/playground.go

@@ -316,6 +316,12 @@ func (p *Playground) handleScaleIn(w io.Writer, pid int) error {
 				p.ticdcs = append(p.ticdcs[:i], p.ticdcs[i+1:]...)
 			}
 		}
+	case spec.ComponentTiProxy:
+		for i := 0; i < len(p.tiproxys); i++ {
+			if p.tiproxys[i].Pid() == pid {
+				p.tiproxys = append(p.tiproxys[:i], p.tiproxys[i+1:]...)
+			}
+		}
 	case spec.ComponentTiKVCDC:
 		for i := 0; i < len(p.tikvCdcs); i++ {
 			if p.tikvCdcs[i].Pid() == pid {
@@ -428,6 +434,8 @@ func (p *Playground) sanitizeComponentConfig(cid string, cfg *instance.Config) e
 		return p.sanitizeConfig(p.bootOptions.Pump, cfg)
 	case spec.ComponentDrainer:
 		return p.sanitizeConfig(p.bootOptions.Drainer, cfg)
+	case spec.ComponentTiProxy:
+		return p.sanitizeConfig(p.bootOptions.TiProxy, cfg)
 	default:
 		return fmt.Errorf("unknown %s in sanitizeConfig", cid)
 	}
@@ -895,7 +903,9 @@ func (p *Playground) bindVersion(comp string, version string) (bindVersion strin
 	default:
 	}
 	if bindVersion == "" {
-		bindVersion = version
+		if version == "nightly" {
+			bindVersion = version
+		}
 	}
 	return
 }

embed/examples/cluster/local.tpl

@@ -10,6 +10,10 @@ global:
   # group is used to specify the group name the user belong to if it's not the same as user.
   group: "{{ .GlobalGroup }}"
   {{- end }}
+  {{- if .GlobalSystemdMode }}
+  # # systemd_mode is used to select whether to use sudo permissions.
+  systemd_mode: "{{ .GlobalSystemdMode }}"
+  {{- end }}
   # # SSH port of servers in the managed cluster.
   ssh_port: {{ .GlobalSSHPort }}
   # # Storage directory for cluster deployment files, startup scripts, and configuration files.
@@ -62,4 +66,4 @@ alertmanager_servers:
  {{- range .AlertManagerServers }}
   - host: {{ . }}
 {{- end }}
-{{ end }}
+{{ end }}

embed/examples/cluster/minimal.yaml

@@ -5,6 +5,8 @@ global:
   user: "tidb"
   # # group is used to specify the group name the user belong to if it's not the same as user.
   # group: "tidb"
+  # # systemd_mode is used to select whether to use sudo permissions. When its value is set to user, there is no need to add global.user to sudoers. The default value is system.
+  # systemd_mode: "system"
   # # SSH port of servers in the managed cluster.
   ssh_port: 22
   # # Storage directory for cluster deployment files, startup scripts, and configuration files.

embed/examples/cluster/multi-dc.yaml

@@ -6,6 +6,8 @@ global:
   # # group is used to specify the group name the user belong to,if it's not the same as user.
   # group: "tidb"
   # # SSH port of servers in the managed cluster.
+  # # systemd_mode is used to select whether to use sudo permissions. When its value is set to user, there is no need to add global.user to sudoers. The default value is system.
+  # systemd_mode: "system"
   ssh_port: 22
   # # Storage directory for cluster deployment files, startup scripts, and configuration files.
   deploy_dir: "/tidb-deploy"

embed/examples/cluster/topology.example.yaml

@@ -6,6 +6,8 @@ global:
   # # group is used to specify the group name the user belong to,if it's not the same as user.
   # group: "tidb"
   # # SSH port of servers in the managed cluster.
+  # # systemd_mode is used to select whether to use sudo permissions. When its value is set to user, there is no need to add global.user to sudoers. The default value is system.
+  # systemd_mode: "system"
   ssh_port: 22
   # # Storage directory for cluster deployment files, startup scripts, and configuration files.
   deploy_dir: "/tidb-deploy"

embed/examples/dm/local.tpl

@@ -11,6 +11,10 @@ global:
   # group is used to specify the group name the user belong to if it's not the same as user.
   group: "{{ .GlobalGroup }}"
   {{- end }}
+  {{- if .GlobalSystemdMode }}
+  # # systemd_mode is used to select whether to use sudo permissions.
+  systemd_mode: "{{ .GlobalSystemdMode }}"
+  {{- end }}
   # # SSH port of servers in the managed cluster.
   ssh_port: {{ .GlobalSSHPort }}
   # # Storage directory for cluster deployment files, startup scripts, and configuration files.
@@ -51,4 +55,4 @@ alertmanager_servers:
  {{- range .AlertManagerServers }}
   - host: {{ . }}
 {{- end }}
-{{ end }}
+{{ end }}

embed/examples/dm/minimal.yaml

@@ -6,6 +6,7 @@
 ---
 global:
   user: "tidb"
+  # systemd_mode: "system"
   ssh_port: 22
   deploy_dir: "/home/tidb/dm/deploy"
   data_dir: "/home/tidb/dm/data"

embed/examples/dm/topology.example.yaml

@@ -3,6 +3,7 @@
 # them if the specific deployment value missing.
 global:
   user: "tidb"
+  # systemd_mode: "system"
   ssh_port: 22
   deploy_dir: "/dm-deploy"
   data_dir: "/dm-data"

embed/templates/config/prometheus.yml.tpl

@@ -376,8 +376,10 @@ scrape_configs:
         target_label: __param_target
       - source_labels: [__param_target]
         target_label: instance
+      {{- if .BlackboxAddr}}
       - target_label: __address__
         replacement: '{{.BlackboxAddr}}'
+      {{- end}}
 {{- range $addr := .BlackboxExporterAddrs}}
   - job_name: "blackbox_exporter_{{$addr}}_icmp"
     scrape_interval: 6s

embed/templates/systemd/system.service.tpl

@@ -21,10 +21,12 @@ LimitCORE={{.LimitCORE}}
 LimitNOFILE=1000000
 LimitSTACK=10485760
 
-{{- if .GrantCapNetRaw}}
+{{- if and .GrantCapNetRaw (eq .SystemdMode "system")}}
 AmbientCapabilities=CAP_NET_RAW
 {{- end}}
+{{- if eq .SystemdMode "system"}}
 User={{.User}}
+{{- end}}
 ExecStart=/bin/bash -c '{{.DeployDir}}/scripts/run_{{.ServiceName}}.sh'
 {{- if eq .ServiceName "prometheus"}}
 ExecReload=/bin/bash -c 'kill -HUP $MAINPID $(pidof {{.DeployDir}}/bin/ng-monitoring-server)'

pkg/cluster/ctxt/context.go

@@ -37,9 +37,9 @@ const (
 
 type (
 	// Executor is the executor interface for TiUP, all tasks will in the end
-	// be passed to a executor and then be actually performed.
+	// be passed to an executor and then be actually performed.
 	Executor interface {
-		// Execute run the command, then return it's stdout and stderr
+		// Execute run the command, then return its stdout and stderr
 		// NOTE: stdin is not supported as it seems we don't need it (for now). If
 		// at some point in the future we need to pass stdin to a command, we'll
 		// need to refactor this function and its implementations.

pkg/cluster/executor/ssh.go

@@ -143,8 +143,8 @@ func (e *EasySSHExecutor) Execute(ctx context.Context, cmd string, sudo bool, ti
 	}
 
 	// set a basic PATH in case it's empty on login
-	cmd = fmt.Sprintf("PATH=$PATH:/bin:/sbin:/usr/bin:/usr/sbin %s", cmd)
-
+	cmd = fmt.Sprintf("PATH=$PATH:/bin:/sbin:/usr/bin:/usr/sbin; %s", cmd)
+	
 	if e.Locale != "" {
 		cmd = fmt.Sprintf("export LANG=%s; %s", e.Locale, cmd)
 	}
@@ -182,6 +182,7 @@ func (e *EasySSHExecutor) Execute(ctx context.Context, cmd string, sudo bool, ti
 					e.Config.Server,
 					color.YellowString(output)))
 		}
+
 		return []byte(stdout), []byte(stderr), baseErr
 	}