Merge pull request #2 from krallin/master

Initial PR

.gitignore

@@ -0,0 +1,2 @@
+supercronic
+vendor

.travis.yml

@@ -0,0 +1,35 @@
+sudo: false
+dist: trusty
+
+language: go
+
+go: "1.8"
+
+install:
+  - curl https://glide.sh/get | sh
+  - make deps
+  - git clone https://github.com/sstephenson/bats.git --branch v0.4.0 --depth 1 "${HOME}/bats"
+  - export "PATH=${PATH}:${HOME}/bats/bin"
+
+jobs:
+  include:
+    - stage: test
+      script:
+        - make fmt && make test
+    - stage: build
+      script:
+        - mkdir -p dist
+        - for arch in amd64 386 arm arm64; do GOARCH="$arch" go build && mv supercronic "dist/supercronic-${arch}"; done
+        - cd dist
+        - ls -lah *
+        - file *
+        - sha1sum *
+        - sha256sum *
+      deploy:
+        provider: releases
+        api_key:
+          secure: "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"
+        file: "dist/*"
+        skip_cleanup: true
+        on:
+          tags: true

LICENSE.md

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2017, Aptible, Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+of the Software, and to permit persons to whom the Software is furnished to do
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

Makefile

@@ -0,0 +1,27 @@
+GOFILES_NOVENDOR = $(shell find . -type f -name '*.go' -not -path "./vendor/*")
+SHELL=/bin/bash
+
+.PHONY: deps
+deps:
+	glide install
+
+.PHONY: build
+build: $(GOFILES)
+	go build
+
+.PHONY: unit
+unit:
+	go test $$(go list ./... | grep -v /vendor/)
+	go vet $$(go list ./... | grep -v /vendor/)
+
+.PHONY: integration
+integration: build
+	bats integration
+
+.PHONY: test
+test: unit integration
+	true
+
+.PHONY: fmt
+fmt:
+	gofmt -l -w ${GOFILES_NOVENDOR}

README.md

@@ -0,0 +1,152 @@
+# Supercronic #
+
+Supercronic is a crontab-compatible job runner, designed specifically to run in
+containers.
+
+
+## Why Supercronic? ##
+
+Crontabs are the lingua franca of job scheduling, but typical server cron
+implementations are ill-suited for container environments:
+
+- They purge their environment before starting jobs. This is an important
+  security feature in multi-user systems, but it breaks a fundamental
+  configuration mechanism for containers.
+- They capture the output from the jobs they run, and often either want to
+  email this output or simply discard it. In a containerized environment,
+  logging task output and errors to `stdout` / `stderr` is often easier to work
+  with.
+- They often don't respond gracefully to `SIGINT` / `SIGTERM`, and may leave
+  running jobs orphaned when signaled. Again, this makes sense in a server
+  environment where `init` will handle the orphan jobs and Cron isn't restarted
+  often anyway, but it's inappropriate in a container environment as it'll
+  result in jobs being forcefully terminated (i.e.  `SIGKILL`'ed) when the
+  container exits.
+- They often try to send their logs to syslog. This conveniently provides
+  centralized logging when a syslog server is running, but with containers,
+  simply logging to `stdout` or `stderr` is preferred.
+
+Finally, they are often very quiet, which makes the above issues difficult to
+understand or debug!
+
+The list could go on, but the fundamental takeaway is this: unlike typical
+server cron implementations, Supercronic tries very hard to do exactly what you
+expect from running `cron` in a container:
+
+- Your environment variables are available in jobs.
+- Job output is logged to `stdout` / `stderr`.
+- `SIGTERM` (or `SIGINT`, which you can deliver via CTRL+C when used
+  interactively) triggers a graceful shutdown
+- Job return codes and schedules are also logged to `stdout` / `stderr`.
+
+## How does it work? ##
+
+- Install Supercronic (see below).
+- Point it at a crontab: `supercronic CRONTAB`.
+- You're done!
+
+
+### Installation
+
+- If you have a `go` toolchain available: `go install github.com/aptible/supercronic`
+- TODO: Docker installation instructions / packaging.
+
+
+## Crontab format ##
+
+Broadly speaking, Supercronic tries to process crontabs just like Vixie cron
+does. In most cases, it should be compatible with your existing crontab.
+
+There are, however, a few exceptions:
+
+- First, Supercronic supports second-resolution schedules: under the hood,
+  Supercronic uses [the `cronexpr` package][cronexpr], so refer to its
+  documentation to know exactly what you can do.
+- Second, Supercronic does not support changing users when running tasks.
+  Again, this is something that hardly makes sense in a cron environment. This
+  means that setting `USER` in your crontab won't have any effect.
+
+Here's an example crontab:
+
+```
+# Run every minute
+*/1 * * * * echo "hello"
+
+# Run every 2 seconds
+*/2 * * * * * * ls 2>/dev/null
+```
+
+
+## Environment variables ##
+
+Just like regular cron, Supercronic lets you specify environment variables in
+your crontab using a `KEY=VALUE` syntax.
+
+However, this is only here for compatibility with existing crontabs, and using
+this feature is generally **not recommended** when using Supercronic.
+
+Indeed, Supercronic does not wipe your environment before running jobs, so if
+you need environment variables to be available when your jobs run, just set
+them before starting Supercronic itself, and your jobs will inherit them
+(unless you've used cron before, this is exactly what you expect).
+
+For example, if you're using Docker, Supercronic
+
+
+## Logging ##
+
+Supercronic provides rich logging, and will let you know exactly what command
+triggered a given message. Here's an example:
+
+```
+$ cat ./my-crontab
+*/5 * * * * * * echo "hello from Supercronic"
+
+$ ./supercronic ./my-crontab
+INFO[2017-07-10T19:40:44+02:00] read crontab: ./my-crontab
+INFO[2017-07-10T19:40:50+02:00] starting                                      iteration=0 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+INFO[2017-07-10T19:40:50+02:00] hello from Supercronic                        channel=stdout iteration=0 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+INFO[2017-07-10T19:40:50+02:00] job succeeded                                 iteration=0 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+INFO[2017-07-10T19:40:55+02:00] starting                                      iteration=1 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+INFO[2017-07-10T19:40:55+02:00] hello from Supercronic                        channel=stdout iteration=1 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+INFO[2017-07-10T19:40:55+02:00] job succeeded                                 iteration=1 job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+```
+
+
+## Debugging ##
+
+If your jobs aren't running, or you'd simply like to double-check your crontab
+syntax, pass the `-debug` flag for more verbose logging:
+
+```
+$ ./supercronic -debug ./my-crontab
+INFO[2017-07-10T19:43:51+02:00] read crontab: ./my-crontab
+DEBU[2017-07-10T19:43:51+02:00] try parse(7): */5 * * * * * * echo "hello from Supercronic"[0:15] = */5 * * * * * *
+DEBU[2017-07-10T19:43:51+02:00] job will run next at 2017-07-10 19:44:00 +0200 CEST  job.command="echo "hello from Supercronic"" job.position=0 job.schedule="*/5 * * * * * *"
+```
+
+
+## Duplicate Jobs ##
+
+Supercronic will wait for a given job to finish before that job is scheduled
+again (some cron implementations do this, others don't). If a job is falling
+behind schedule (i.e. it's taking too long to finish), Supercronic will warn
+you.
+
+Here is an example:
+
+```
+# Sleep for 2 seconds every second. This will take too long.
+* * * * * * * sleep 2
+
+$ ./supercronic ./my-crontab
+INFO[2017-07-11T12:24:25+02:00] read crontab: foo
+INFO[2017-07-11T12:24:27+02:00] starting                                      iteration=0 job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+INFO[2017-07-11T12:24:29+02:00] job succeeded                                 iteration=0 job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+WARN[2017-07-11T12:24:29+02:00] job took too long to run: it should have started 1.009438854s ago  job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+INFO[2017-07-11T12:24:30+02:00] starting                                      iteration=1 job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+INFO[2017-07-11T12:24:32+02:00] job succeeded                                 iteration=1 job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+WARN[2017-07-11T12:24:32+02:00] job took too long to run: it should have started 1.014474099s ago  job.command="sleep 2" job.position=0 job.schedule="* * * * * * *"
+```
+
+  [cronexpr]: https://github.com/gorhill/cronexpr

cron/cron.go

@@ -0,0 +1,139 @@
+package cron
+
+import (
+	"bufio"
+	"fmt"
+	"github.com/aptible/supercronic/crontab"
+	"github.com/sirupsen/logrus"
+	"io"
+	"os"
+	"os/exec"
+	"strings"
+	"sync"
+	"syscall"
+	"time"
+)
+
+func startReaderDrain(wg *sync.WaitGroup, readerLogger *logrus.Entry, reader io.Reader) {
+	wg.Add(1)
+
+	go func() {
+		defer wg.Done()
+
+		scanner := bufio.NewScanner(reader)
+
+		for scanner.Scan() {
+			readerLogger.Info(scanner.Text())
+		}
+
+		if err := scanner.Err(); err != nil {
+			// The underlying reader might get closed by e.g. Wait(), or
+			// even the process we're starting, so we don't log EOF-like
+			// errors
+			if strings.Contains(err.Error(), os.ErrClosed.Error()) {
+				return
+			}
+
+			readerLogger.Error(err)
+		}
+	}()
+}
+
+func runJob(context *crontab.Context, command string, jobLogger *logrus.Entry) error {
+	jobLogger.Info("starting")
+
+	cmd := exec.Command(context.Shell, "-c", command)
+
+	// Run in a separate process group so that in interactive usage, CTRL+C
+	// stops supercronic, not the children threads.
+	cmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
+
+	env := os.Environ()
+	for k, v := range context.Environ {
+		env = append(env, fmt.Sprintf("%s=%s", k, v))
+	}
+	cmd.Env = env
+
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		return err
+	}
+
+	stderr, err := cmd.StderrPipe()
+	if err != nil {
+		return err
+	}
+
+	if err := cmd.Start(); err != nil {
+		return err
+	}
+
+	var wg sync.WaitGroup
+
+	stdoutLogger := jobLogger.WithFields(logrus.Fields{"channel": "stdout"})
+	startReaderDrain(&wg, stdoutLogger, stdout)
+
+	stderrLogger := jobLogger.WithFields(logrus.Fields{"channel": "stderr"})
+	startReaderDrain(&wg, stderrLogger, stderr)
+
+	wg.Wait()
+
+	if err := cmd.Wait(); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func StartJob(wg *sync.WaitGroup, context *crontab.Context, job *crontab.Job, exitChan chan interface{}) {
+	wg.Add(1)
+
+	go func() {
+		defer wg.Done()
+
+		cronLogger := logrus.WithFields(logrus.Fields{
+			"job.schedule": job.Schedule,
+			"job.command":  job.Command,
+			"job.position": job.Position,
+		})
+
+		var cronIteration uint64 = 0
+		nextRun := job.Expression.Next(time.Now())
+
+		// NOTE: this (intentionally) does not run multiple instances of the
+		// job concurrently
+		for {
+			nextRun = job.Expression.Next(nextRun)
+			cronLogger.Debugf("job will run next at %v", nextRun)
+
+			delay := nextRun.Sub(time.Now())
+			if delay < 0 {
+				cronLogger.Warningf("job took too long to run: it should have started %v ago", -delay)
+				nextRun = time.Now()
+				continue
+			}
+
+			select {
+			case <-exitChan:
+				cronLogger.Debug("shutting down")
+				return
+			case <-time.After(delay):
+				// Proceed normally
+			}
+
+			jobLogger := cronLogger.WithFields(logrus.Fields{
+				"iteration": cronIteration,
+			})
+
+			err := runJob(context, job.Command, jobLogger)
+
+			if err == nil {
+				jobLogger.Info("job succeeded")
+			} else {
+				jobLogger.Error(err)
+			}
+
+			cronIteration++
+		}
+	}()
+}