chore(misc/loop): Setup the portal loop infra (#1400)

Portal loop infrastructure code.

Co-authored-by: Morgan <git@howl.moe>
Co-authored-by: Morgan <morgan@morganbaz.com>
Co-authored-by: deelawn <dboltz03@gmail.com>
Co-authored-by: Guilhem Fanton <8671905+gfanton@users.noreply.github.com>
Co-authored-by: Leon Hudak <33522493+leohhhn@users.noreply.github.com>
Co-authored-by: Miloš Živković <milos.zivkovic@tendermint.com>
Co-authored-by: Lee ByeongJun <lbj199874@gmail.com>
Co-authored-by: Kazaï <149690535+kazai777@users.noreply.github.com>
Co-authored-by: Manfred Touron <94029+moul@users.noreply.github.com>

.github/workflows/portal-loop.yml

@@ -0,0 +1,48 @@
+name: portal-loop
+
+on:
+  push:
+    paths:
+      - misc/loop
+      - .github/workflows/portal-loop.yml
+    branches:
+      - "master"
+      - "ops/portal-loop"
+    tags:
+      - "v*"
+
+permissions:
+  contents: read
+  packages: write
+
+jobs:
+  portal-loop:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+
+      - name: Login to GitHub Container Registry
+        uses: docker/login-action@v2
+        with:
+          registry: ghcr.io
+          username: ${{ github.repository_owner }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Docker metadata portalloopd
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository }}/portalloopd
+          tags: |
+            type=raw,value=latest
+            type=semver,pattern=v{{version}}
+
+      - name: Build and push
+        uses: docker/build-push-action@v4
+        with:
+          context: ./misc/loop
+          target: portalloopd
+          push: ${{ github.event_name != 'pull_request' }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

misc/loop/.gitignore

@@ -0,0 +1,3 @@
+/portalloopd
+/backups
+/traefik/letsencrypt

misc/loop/Dockerfile

@@ -0,0 +1,19 @@
+FROM golang:alpine AS builder
+
+COPY . /go/src/github.com/gnolang/gno/misc/loop
+
+WORKDIR /go/src/github.com/gnolang/gno/misc/loop
+
+RUN --mount=type=cache,target=/root/.cache/go-build \
+    --mount=type=cache,target=/root/go/pkg/mod \
+    go build -o /build/portalloopd ./cmd
+
+# Final image for portalloopd
+FROM docker AS portalloopd
+
+RUN apk add bash curl jq
+
+COPY --from=builder /build/portalloopd /usr/bin/portalloopd
+
+ENTRYPOINT [ "/usr/bin/portalloopd" ]
+CMD [ "serve" ]

misc/loop/Makefile

@@ -1,44 +1,22 @@
-# The startup delay (waits until the node is "ready")
-DELAY ?= 10 # seconds
-# The temporary backup file for transactions
-BACKUP_FILE ?= $(abspath ./txs_backup.log)
-# The entire txs history across all iterations
-HISTORY_OUTPUT := $(abspath ./txs_history.log)
+all: docker.start
 
-# The gnoland binary
-gnoland_bin := go run github.com/gnolang/gno/gno.land/cmd/gnoland
-# The tx archive binary
-tx_bin := go run github.com/gnolang/tx-archive/cmd
+docker.start: # Start the portal loop
+	docker compose up -d
 
-# The relative gno.land directory
-gnoland_dir := $(abspath ../../gno.land)
+docker.stop: # Stop the portal loop
+	docker compose down
+	docker rm -f $(docker ps -aq --filter "label=the-portal-loop")
 
-all: loop
+docker.build: # (re)Build snapshotter image
+	docker compose build
 
-start.gnoland:
-	cd $(gnoland_dir) && $(gnoland_bin) start -skip-failing-genesis-txs -genesis-txs-file $(HISTORY_OUTPUT)
-clean.gnoland:
-	make -C $(gnoland_dir) fclean
-.PHONY: start.gnoland clean.gnoland
+docker.pull: # Pull new images to update versions
+	docker compose pull
 
-# Starts the backup service
-# and backs up transactions into a file
-# that is wiped on every loop
-tx.backup:
-	sleep $(DELAY)
-	$(tx_bin) backup -legacy -watch -overwrite -output-path "$(BACKUP_FILE)"
-.PHONY: tx.backup
+portalloopd.bash: # Get a bash command inside of the portalloopd container
+	docker compose exec portalloopd bash
 
-# Saves the history from previous iterations into
-# a temporary transactions log
-save.history:
-	@test -e $(BACKUP_FILE) || (echo "No existing backup file not found: '$(BACKUP_FILE)'"; exit 1)
-	cat $(BACKUP_FILE) >> $(HISTORY_OUTPUT)
-.PHONY: save.history
+switch: portalloopd.switch
 
-loop: clean.gnoland
-	# backup history, if needed
-	$(MAKE) save.history || true
-	# run our dev loop
-	./run_loop.sh
-.PHONY: loop
+portalloopd.switch: # Force switch the portal loop with latest image
+	docker compose exec portalloopd switch

misc/loop/README.md

@@ -0,0 +1,42 @@
+# The portal loop :infinity: 
+
+## What is it?
+
+It's a Gnoland node that aim to run with always the latest version of gno and never loose transactions history.
+
+For more information, see issue on github [gnolang/gno#1239](https://github.com/gnolang/gno/issues/1239)
+
+
+## How to use
+
+Start the loop with:
+
+```sh
+$ docker compose up -d
+
+# or using the Makefile
+
+$ make
+```
+
+The [`portalloopd`](./cmd/portalloopd) binary is starting inside of the docker container `portalloopd`
+
+This script is doing:
+
+- Setup the current portal-loop in read only mode
+- Pull the latest version of [ghcr.io/gnolang/gno]()
+- Backup the txs using [gnolang/tx-archive](https://github.com/gnolang/tx-archive)
+- Start a new docker container with the backups files
+- Changing the proxy (traefik) to redirect to the new portal loop
+- Unlock read only mode
+- Stop the previous loop
+
+### Makefile helper
+
+You can find a [Makefile](./Makefile) to help you interact with the portal loop
+
+- Force switch of the portal loop with a new version
+
+```bash
+make portalloopd.switch
+```

misc/loop/cmd/cmd_backup.go

@@ -0,0 +1,85 @@
+package main
+
+import (
+	"context"
+	"flag"
+	"os"
+
+	"github.com/docker/docker/client"
+	"github.com/gnolang/gno/tm2/pkg/commands"
+)
+
+type backupCfg struct {
+	rpcAddr        string
+	traefikGnoFile string
+	backupDir      string
+	hostPWD        string
+}
+
+func (c *backupCfg) RegisterFlags(fs *flag.FlagSet) {
+	if os.Getenv("HOST_PWD") == "" {
+		os.Setenv("HOST_PWD", os.Getenv("PWD"))
+	}
+
+	if os.Getenv("BACKUP_DIR") == "" {
+		os.Setenv("BACKUP_DIR", "./backups")
+	}
+
+	if os.Getenv("RPC_URL") == "" {
+		os.Setenv("RPC_URL", "http://rpc.portal.gno.local:26657")
+	}
+
+	if os.Getenv("PROM_ADDR") == "" {
+		os.Setenv("PROM_ADDR", ":9090")
+	}
+
+	if os.Getenv("TRAEFIK_GNO_FILE") == "" {
+		os.Setenv("TRAEFIK_GNO_FILE", "./traefik/gno.yml")
+	}
+
+	fs.StringVar(&c.rpcAddr, "rpc", os.Getenv("RPC_URL"), "tendermint rpc url")
+	fs.StringVar(&c.traefikGnoFile, "traefik-gno-file", os.Getenv("TRAEFIK_GNO_FILE"), "traefik gno file")
+	fs.StringVar(&c.backupDir, "backup-dir", os.Getenv("BACKUP_DIR"), "backup directory")
+	fs.StringVar(&c.hostPWD, "pwd", os.Getenv("HOST_PWD"), "host pwd (for docker usage)")
+}
+
+func newBackupCmd(io commands.IO) *commands.Command {
+	cfg := &backupCfg{}
+
+	return commands.NewCommand(
+		commands.Metadata{
+			Name:       "backup",
+			ShortUsage: "backup [flags]",
+		},
+		cfg,
+		func(ctx context.Context, _ []string) error {
+			return execBackup(ctx, cfg)
+		},
+	)
+}
+
+func execBackup(ctx context.Context, cfg *backupCfg) error {
+	dockerClient, err := client.NewEnvClient()
+	if err != nil {
+		return err
+	}
+
+	portalLoop := &snapshotter{}
+
+	portalLoop, err = NewSnapshotter(dockerClient, config{
+		backupDir:      cfg.backupDir,
+		rpcAddr:        cfg.rpcAddr,
+		hostPWD:        cfg.hostPWD,
+		traefikGnoFile: cfg.traefikGnoFile,
+	})
+	if err != nil {
+		return err
+	}
+
+	err = StartPortalLoop(ctx, portalLoop, false)
+	if err != nil {
+		return err
+	}
+
+	return portalLoop.backupTXs(ctx, portalLoop.url)
+}

misc/loop/cmd/cmd_serve.go

@@ -0,0 +1,116 @@
+package main
+
+import (
+	"context"
+	"flag"
+	"net/http"
+	"os"
+	"time"
+
+	"github.com/docker/docker/client"
+	"github.com/gnolang/gno/tm2/pkg/commands"
+	"github.com/prometheus/client_golang/prometheus/promhttp"
+	"github.com/sirupsen/logrus"
+)
+
+type serveCfg struct {
+	rpcAddr        string
+	traefikGnoFile string
+	backupDir      string
+	hostPWD        string
+}
+
+type serveService struct {
+	cfg serveCfg
+
+	// TODO(albttx): put getter on it with RMutex
+	portalLoop *snapshotter
+
+	portalLoopURL string
+}
+
+func (c *serveCfg) RegisterFlags(fs *flag.FlagSet) {
+	if os.Getenv("HOST_PWD") == "" {
+		os.Setenv("HOST_PWD", os.Getenv("PWD"))
+	}
+
+	if os.Getenv("BACKUP_DIR") == "" {
+		os.Setenv("BACKUP_DIR", "./backups")
+	}
+
+	if os.Getenv("RPC_URL") == "" {
+		os.Setenv("RPC_URL", "http://rpc.portal.gno.local:26657")
+	}
+
+	if os.Getenv("PROM_ADDR") == "" {
+		os.Setenv("PROM_ADDR", ":9090")
+	}
+
+	if os.Getenv("TRAEFIK_GNO_FILE") == "" {
+		os.Setenv("TRAEFIK_GNO_FILE", "./traefik/gno.yml")
+	}
+
+	fs.StringVar(&c.rpcAddr, "rpc", os.Getenv("RPC_URL"), "tendermint rpc url")
+	fs.StringVar(&c.traefikGnoFile, "traefik-gno-file", os.Getenv("TRAEFIK_GNO_FILE"), "traefik gno file")
+	fs.StringVar(&c.backupDir, "backup-dir", os.Getenv("BACKUP_DIR"), "backup directory")
+	fs.StringVar(&c.hostPWD, "pwd", os.Getenv("HOST_PWD"), "host pwd (for docker usage)")
+}
+
+func newServeCmd(io commands.IO) *commands.Command {
+	cfg := &serveCfg{}
+
+	return commands.NewCommand(
+		commands.Metadata{
+			Name:       "serve",
+			ShortUsage: "serve [flags]",
+		},
+		cfg,
+		func(ctx context.Context, args []string) error {
+			return execServe(ctx, cfg, args)
+		},
+	)
+}
+
+func execServe(ctx context.Context, cfg *serveCfg, args []string) error {
+	dockerClient, err := client.NewEnvClient()
+	if err != nil {
+		return err
+	}
+
+	portalLoop := &snapshotter{}
+
+	// Serve monitoring
+	go func() {
+		s := &monitoringService{
+			portalLoop: portalLoop,
+		}
+
+		for portalLoop.url == "" {
+			time.Sleep(time.Second * 1)
+		}
+
+		go s.recordMetrics()
+
+		http.Handle("/metrics", promhttp.Handler())
+		http.ListenAndServe(os.Getenv("PROM_ADDR"), nil)
+	}()
+
+	// the loop
+	for {
+		portalLoop, err = NewSnapshotter(dockerClient, config{
+			backupDir:      cfg.backupDir,
+			rpcAddr:        cfg.rpcAddr,
+			hostPWD:        cfg.hostPWD,
+			traefikGnoFile: cfg.traefikGnoFile,
+		})
+		if err != nil {
+			return err
+		}
+
+		err = StartPortalLoop(ctx, portalLoop, false)
+		if err != nil {
+			logrus.WithError(err).Error()
+		}
+		time.Sleep(time.Second * 120)
+	}
+}