puma 3.11.2 → 6.0.0
Sign up to get free protection for your applications and to get access to all the features.
Potentially problematic release.
This version of puma might be problematic. Click here for more details.
- checksums.yaml +5 -5
- data/History.md +1708 -422
- data/LICENSE +23 -20
- data/README.md +190 -64
- data/bin/puma-wild +3 -9
- data/docs/architecture.md +59 -21
- data/docs/compile_options.md +55 -0
- data/docs/deployment.md +69 -58
- data/docs/fork_worker.md +31 -0
- data/docs/images/puma-connection-flow-no-reactor.png +0 -0
- data/docs/images/puma-connection-flow.png +0 -0
- data/docs/images/puma-general-arch.png +0 -0
- data/docs/jungle/README.md +9 -0
- data/docs/jungle/rc.d/README.md +74 -0
- data/docs/jungle/rc.d/puma +61 -0
- data/docs/jungle/rc.d/puma.conf +10 -0
- data/docs/kubernetes.md +66 -0
- data/docs/nginx.md +1 -1
- data/docs/plugins.md +22 -12
- data/docs/rails_dev_mode.md +28 -0
- data/docs/restart.md +47 -22
- data/docs/signals.md +13 -11
- data/docs/stats.md +142 -0
- data/docs/systemd.md +100 -115
- data/docs/testing_benchmarks_local_files.md +150 -0
- data/docs/testing_test_rackup_ci_files.md +36 -0
- data/ext/puma_http11/PumaHttp11Service.java +2 -2
- data/ext/puma_http11/ext_help.h +1 -1
- data/ext/puma_http11/extconf.rb +61 -3
- data/ext/puma_http11/http11_parser.c +106 -118
- data/ext/puma_http11/http11_parser.h +2 -2
- data/ext/puma_http11/http11_parser.java.rl +22 -38
- data/ext/puma_http11/http11_parser.rl +6 -4
- data/ext/puma_http11/http11_parser_common.rl +6 -6
- data/ext/puma_http11/mini_ssl.c +376 -93
- data/ext/puma_http11/no_ssl/PumaHttp11Service.java +15 -0
- data/ext/puma_http11/org/jruby/puma/Http11.java +108 -116
- data/ext/puma_http11/org/jruby/puma/Http11Parser.java +84 -99
- data/ext/puma_http11/org/jruby/puma/MiniSSL.java +250 -88
- data/ext/puma_http11/puma_http11.c +49 -57
- data/lib/puma/app/status.rb +71 -49
- data/lib/puma/binder.rb +243 -148
- data/lib/puma/cli.rb +50 -35
- data/lib/puma/client.rb +369 -232
- data/lib/puma/cluster/worker.rb +175 -0
- data/lib/puma/cluster/worker_handle.rb +97 -0
- data/lib/puma/cluster.rb +268 -235
- data/lib/puma/commonlogger.rb +4 -2
- data/lib/puma/configuration.rb +116 -88
- data/lib/puma/const.rb +49 -30
- data/lib/puma/control_cli.rb +124 -77
- data/lib/puma/detect.rb +33 -2
- data/lib/puma/dsl.rb +685 -138
- data/lib/puma/error_logger.rb +112 -0
- data/lib/puma/events.rb +17 -111
- data/lib/puma/io_buffer.rb +34 -5
- data/lib/puma/jruby_restart.rb +4 -59
- data/lib/puma/json_serialization.rb +96 -0
- data/lib/puma/launcher/bundle_pruner.rb +104 -0
- data/lib/puma/launcher.rb +197 -130
- data/lib/puma/log_writer.rb +137 -0
- data/lib/puma/minissl/context_builder.rb +92 -0
- data/lib/puma/minissl.rb +256 -70
- data/lib/puma/null_io.rb +20 -1
- data/lib/puma/plugin/tmp_restart.rb +3 -1
- data/lib/puma/plugin.rb +9 -13
- data/lib/puma/rack/builder.rb +8 -9
- data/lib/puma/rack/urlmap.rb +2 -0
- data/lib/puma/rack_default.rb +3 -1
- data/lib/puma/reactor.rb +90 -187
- data/lib/puma/request.rb +607 -0
- data/lib/puma/runner.rb +94 -71
- data/lib/puma/server.rb +336 -703
- data/lib/puma/single.rb +27 -72
- data/lib/puma/state_file.rb +46 -7
- data/lib/puma/systemd.rb +47 -0
- data/lib/puma/thread_pool.rb +185 -91
- data/lib/puma/util.rb +23 -10
- data/lib/puma.rb +68 -3
- data/lib/rack/handler/puma.rb +17 -14
- data/tools/Dockerfile +16 -0
- data/tools/trickletest.rb +0 -1
- metadata +53 -30
- data/ext/puma_http11/io_buffer.c +0 -155
- data/lib/puma/accept_nonblock.rb +0 -23
- data/lib/puma/compat.rb +0 -14
- data/lib/puma/convenient.rb +0 -23
- data/lib/puma/daemon_ext.rb +0 -31
- data/lib/puma/delegation.rb +0 -11
- data/lib/puma/java_io_buffer.rb +0 -45
- data/lib/puma/rack/backports/uri/common_193.rb +0 -33
- data/lib/puma/tcp_logger.rb +0 -39
- data/tools/jungle/README.md +0 -13
- data/tools/jungle/init.d/README.md +0 -59
- data/tools/jungle/init.d/puma +0 -421
- data/tools/jungle/init.d/run-puma +0 -18
- data/tools/jungle/upstart/README.md +0 -61
- data/tools/jungle/upstart/puma-manager.conf +0 -31
- data/tools/jungle/upstart/puma.conf +0 -69
data/docs/restart.md
CHANGED
@@ -1,39 +1,64 @@
|
|
1
|
-
|
1
|
+
Puma provides three distinct kinds of restart operations, each for different use cases. This document describes "hot restarts" and "phased restarts." The third kind of restart operation is called "refork" and is described in the documentation for [`fork_worker`](fork_worker.md).
|
2
2
|
|
3
|
-
|
3
|
+
## Hot restart
|
4
4
|
|
5
|
-
|
6
|
-
* Send the `puma` process the `SIGUSR1` signal (rolling restart, cluster mode only)
|
7
|
-
* Use the status server and issue `/restart`
|
5
|
+
To perform a "hot" restart, Puma performs an `exec` operation to start the process up again, so no memory is shared between the old process and the new process. As a result, it is safe to issue a restart at any place where you would manually stop Puma and start it again. In particular, it is safe to upgrade Puma itself using a hot restart.
|
8
6
|
|
9
|
-
|
7
|
+
If the new process is unable to load, it will simply exit. You should therefore run Puma under a process monitor when using it in production.
|
10
8
|
|
11
|
-
|
9
|
+
### How-to
|
12
10
|
|
13
|
-
|
11
|
+
Any of the following will cause a Puma server to perform a hot restart:
|
14
12
|
|
15
|
-
|
13
|
+
* Send the `puma` process the `SIGUSR2` signal
|
14
|
+
* Issue a `GET` request to the Puma status/control server with the path `/restart`
|
15
|
+
* Issue `pumactl restart` (this uses the control server method if available, otherwise sends the `SIGUSR2` signal to the process)
|
16
16
|
|
17
|
-
|
17
|
+
### Supported configurations
|
18
18
|
|
19
|
-
|
19
|
+
* Works in cluster mode and single mode
|
20
|
+
* Supported on all platforms
|
20
21
|
|
21
|
-
|
22
|
+
### Client experience
|
22
23
|
|
23
|
-
|
24
|
+
* All platforms: clients with an in-flight request are served responses before the connection is closed gracefully. Puma gracefully disconnects any idle HTTP persistent connections before restarting.
|
25
|
+
* On MRI or TruffleRuby on Linux and BSD: Clients who connect just before the server restarts may experience increased latency while the server stops and starts again, but their connections will not be closed prematurely.
|
26
|
+
* On Windows and JRuby: Clients who connect just before a restart may experience "connection reset" errors.
|
24
27
|
|
25
|
-
###
|
28
|
+
### Additional notes
|
26
29
|
|
27
|
-
|
30
|
+
* Only one version of the application is running at a time.
|
31
|
+
* `on_restart` is invoked just before the server shuts down. This can be used to clean up resources (like long-lived database connections) gracefully. Since Ruby 2.0, it is not typically necessary to explicitly close file descriptors on restart. This is because any file descriptor opened by Ruby will have the `FD_CLOEXEC` flag set, meaning that file descriptors are closed on `exec`. `on_restart` is useful, though, if your application needs to perform any more graceful protocol-specific shutdown procedures before closing connections.
|
28
32
|
|
29
|
-
|
30
|
-
# config/puma.rb
|
33
|
+
## Phased restart
|
31
34
|
|
32
|
-
|
33
|
-
```
|
35
|
+
Phased restarts replace all running workers in a Puma cluster. This is a useful way to upgrade the application that Puma is serving gracefully. A phased restart works by first killing an old worker, then starting a new worker, waiting until the new worker has successfully started before proceeding to the next worker. This process continues until all workers are replaced. The master process is not restarted.
|
34
36
|
|
35
|
-
###
|
37
|
+
### How-to
|
36
38
|
|
37
|
-
|
39
|
+
Any of the following will cause a Puma server to perform a phased restart:
|
38
40
|
|
39
|
-
|
41
|
+
* Send the `puma` process the `SIGUSR1` signal
|
42
|
+
* Issue a `GET` request to the Puma status/control server with the path `/phased-restart`
|
43
|
+
* Issue `pumactl phased-restart` (this uses the control server method if available, otherwise sends the `SIGUSR1` signal to the process)
|
44
|
+
|
45
|
+
### Supported configurations
|
46
|
+
|
47
|
+
* Works in cluster mode only
|
48
|
+
* To support upgrading the application that Puma is serving, ensure `prune_bundler` is enabled and that `preload_app!` is disabled
|
49
|
+
* Supported on all platforms where cluster mode is supported
|
50
|
+
|
51
|
+
### Client experience
|
52
|
+
|
53
|
+
* In-flight requests are always served responses before the connection is closed gracefully
|
54
|
+
* Idle persistent connections are gracefully disconnected
|
55
|
+
* New connections are not lost, and clients will not experience any increase in latency (as long as the number of configured workers is greater than one)
|
56
|
+
|
57
|
+
### Additional notes
|
58
|
+
|
59
|
+
* When a phased restart begins, the Puma master process changes its current working directory to the directory specified by the `directory` option. If `directory` is set to symlink, this is automatically re-evaluated, so this mechanism can be used to upgrade the application.
|
60
|
+
* On a single server, it's possible that two versions of the application are running concurrently during a phased restart.
|
61
|
+
* `on_restart` is not invoked
|
62
|
+
* Phased restarts can be slow for Puma clusters with many workers. Hot restarts often complete more quickly, but at the cost of increased latency during the restart.
|
63
|
+
* Phased restarts cannot be used to upgrade any gems loaded by the Puma master process, including `puma` itself, anything in `extra_runtime_dependencies`, or dependencies thereof. Upgrading other gems is safe.
|
64
|
+
* If you remove the gems from old releases as part of your deployment strategy, there are additional considerations. Do not put any gems into `extra_runtime_dependencies` that have native extensions or have dependencies that have native extensions (one common example is `puma_worker_killer` and its dependency on `ffi`). Workers will fail on boot during a phased restart. The underlying issue is recorded in [an issue on the rubygems project](https://github.com/rubygems/rubygems/issues/4004). Hot restarts are your only option here if you need these dependencies.
|
data/docs/signals.md
CHANGED
@@ -1,8 +1,8 @@
|
|
1
|
-
The [unix signal](
|
1
|
+
The [unix signal](https://en.wikipedia.org/wiki/Unix_signal) is a method of sending messages between [processes](https://en.wikipedia.org/wiki/Process_(computing)). When a signal is sent, the operating system interrupts the target process's normal flow of execution. There are standard signals that are used to stop a process, but there are also custom signals that can be used for other purposes. This document is an attempt to list all supported signals that Puma will respond to. In general, signals need only be sent to the master process of a cluster.
|
2
2
|
|
3
3
|
## Sending Signals
|
4
4
|
|
5
|
-
If you are new to signals it can be
|
5
|
+
If you are new to signals, it can be helpful to see how they are used. When a process starts in a *nix-like operating system, it will have a [PID - or process identifier](https://en.wikipedia.org/wiki/Process_identifier) that can be used to send signals to the process. For demonstration, we will create an infinitely running process by tailing a file:
|
6
6
|
|
7
7
|
```sh
|
8
8
|
$ echo "foo" >> my.log
|
@@ -10,24 +10,24 @@ $ irb
|
|
10
10
|
> pid = Process.spawn 'tail -f my.log'
|
11
11
|
```
|
12
12
|
|
13
|
-
From here we can see that the tail process is running by using the `ps` command:
|
13
|
+
From here, we can see that the tail process is running by using the `ps` command:
|
14
14
|
|
15
15
|
```sh
|
16
16
|
$ ps aux | grep tail
|
17
17
|
schneems 87152 0.0 0.0 2432772 492 s032 S+ 12:46PM 0:00.00 tail -f my.log
|
18
18
|
```
|
19
19
|
|
20
|
-
You can send a signal in Ruby using the [Process module](
|
20
|
+
You can send a signal in Ruby using the [Process module](https://www.ruby-doc.org/core-2.1.1/Process.html#kill-method):
|
21
21
|
|
22
22
|
```
|
23
23
|
$ irb
|
24
24
|
> puts pid
|
25
25
|
=> 87152
|
26
|
-
Process.detach(pid) #
|
26
|
+
Process.detach(pid) # https://ruby-doc.org/core-2.1.1/Process.html#method-c-detach
|
27
27
|
Process.kill("TERM", pid)
|
28
28
|
```
|
29
29
|
|
30
|
-
Now you will see via `ps` that there is no more `tail` process. Sometimes when referring to signals the `SIG` prefix will be used
|
30
|
+
Now you will see via `ps` that there is no more `tail` process. Sometimes when referring to signals, the `SIG` prefix will be used. For example, `SIGTERM` is equivalent to sending `TERM` via `Process.kill`.
|
31
31
|
|
32
32
|
## Puma Signals
|
33
33
|
|
@@ -35,12 +35,14 @@ Puma cluster responds to these signals:
|
|
35
35
|
|
36
36
|
- `TTIN` increment the worker count by 1
|
37
37
|
- `TTOU` decrement the worker count by 1
|
38
|
-
- `TERM` send `TERM` to worker.
|
39
|
-
- `USR2` restart workers. This also reloads
|
40
|
-
- `USR1` restart workers in phases, a rolling restart. This will not reload configuration file.
|
41
|
-
- `HUP`
|
42
|
-
- `INT` equivalent of sending Ctrl-C to cluster.
|
38
|
+
- `TERM` send `TERM` to worker. The worker will attempt to finish then exit.
|
39
|
+
- `USR2` restart workers. This also reloads the Puma configuration file, if there is one.
|
40
|
+
- `USR1` restart workers in phases, a rolling restart. This will not reload the configuration file.
|
41
|
+
- `HUP ` reopen log files defined in stdout_redirect configuration parameter. If there is no stdout_redirect option provided, it will behave like `INT`
|
42
|
+
- `INT ` equivalent of sending Ctrl-C to cluster. Puma will attempt to finish then exit.
|
43
43
|
- `CHLD`
|
44
|
+
- `URG ` refork workers in phases from worker 0 if `fork_workers` option is enabled.
|
45
|
+
- `INFO` print backtraces of all puma threads
|
44
46
|
|
45
47
|
## Callbacks order in case of different signals
|
46
48
|
|
data/docs/stats.md
ADDED
@@ -0,0 +1,142 @@
|
|
1
|
+
## Accessing stats
|
2
|
+
|
3
|
+
Stats can be accessed in two ways:
|
4
|
+
|
5
|
+
### control server
|
6
|
+
|
7
|
+
`$ pumactl stats` or `GET /stats`
|
8
|
+
|
9
|
+
[Read more about `pumactl` and the control server in the README.](https://github.com/puma/puma#controlstatus-server).
|
10
|
+
|
11
|
+
### Puma.stats
|
12
|
+
|
13
|
+
`Puma.stats` produces a JSON string. `Puma.stats_hash` produces a ruby hash.
|
14
|
+
|
15
|
+
#### in single mode
|
16
|
+
|
17
|
+
Invoke `Puma.stats` anywhere in runtime, e.g. in a rails initializer:
|
18
|
+
|
19
|
+
```ruby
|
20
|
+
# config/initializers/puma_stats.rb
|
21
|
+
|
22
|
+
Thread.new do
|
23
|
+
loop do
|
24
|
+
sleep 30
|
25
|
+
puts Puma.stats
|
26
|
+
end
|
27
|
+
end
|
28
|
+
```
|
29
|
+
|
30
|
+
#### in cluster mode
|
31
|
+
|
32
|
+
Invoke `Puma.stats` from the master process
|
33
|
+
|
34
|
+
```ruby
|
35
|
+
# config/puma.rb
|
36
|
+
|
37
|
+
before_fork do
|
38
|
+
Thread.new do
|
39
|
+
loop do
|
40
|
+
puts Puma.stats
|
41
|
+
sleep 30
|
42
|
+
end
|
43
|
+
end
|
44
|
+
end
|
45
|
+
```
|
46
|
+
|
47
|
+
|
48
|
+
## Explanation of stats
|
49
|
+
|
50
|
+
`Puma.stats` returns different information and a different structure depending on if Puma is in single vs. cluster mode. There is one top-level attribute that is common to both modes:
|
51
|
+
|
52
|
+
* started_at: when Puma was started
|
53
|
+
|
54
|
+
### single mode and individual workers in cluster mode
|
55
|
+
|
56
|
+
When Puma runs in single mode, these stats are available at the top level. When Puma runs in cluster mode, these stats are available within the `worker_status` array in a hash labeled `last_status`, in an array of hashes where one hash represents each worker.
|
57
|
+
|
58
|
+
* backlog: requests that are waiting for an available thread to be available. if this is above 0, you need more capacity [always true?]
|
59
|
+
* running: how many threads are running
|
60
|
+
* pool_capacity: the number of requests that the server is capable of taking right now. For example, if the number is 5, then it means there are 5 threads sitting idle ready to take a request. If one request comes in, then the value would be 4 until it finishes processing. If the minimum threads allowed is zero, this number will still have a maximum value of the maximum threads allowed.
|
61
|
+
* max_threads: the maximum number of threads Puma is configured to spool per worker
|
62
|
+
* requests_count: the number of requests this worker has served since starting
|
63
|
+
|
64
|
+
|
65
|
+
### cluster mode
|
66
|
+
|
67
|
+
* phase: which phase of restart the process is in, during [phased restart](https://github.com/puma/puma/blob/master/docs/restart.md)
|
68
|
+
* workers: ??
|
69
|
+
* booted_workers: how many workers currently running?
|
70
|
+
* old_workers: ??
|
71
|
+
* worker_status: array of hashes of info for each worker (see below)
|
72
|
+
|
73
|
+
### worker status
|
74
|
+
|
75
|
+
* started_at: when the worker started
|
76
|
+
* pid: the process id of the worker process
|
77
|
+
* index: each worker gets a number. if Puma is configured to have 3 workers, then this will be 0, 1, or 2
|
78
|
+
* booted: if it's done booting [?]
|
79
|
+
* last_checkin: Last time the worker responded to the master process' heartbeat check.
|
80
|
+
* last_status: a hash of info about the worker's state handling requests. See the explanation for this in "single mode and individual workers in cluster mode" section above.
|
81
|
+
|
82
|
+
|
83
|
+
## Examples
|
84
|
+
|
85
|
+
Here are two example stats hashes produced by `Puma.stats`:
|
86
|
+
|
87
|
+
### single
|
88
|
+
|
89
|
+
```json
|
90
|
+
{
|
91
|
+
"started_at": "2021-01-14T07:12:35Z",
|
92
|
+
"backlog": 0,
|
93
|
+
"running": 5,
|
94
|
+
"pool_capacity": 5,
|
95
|
+
"max_threads": 5,
|
96
|
+
"requests_count": 3
|
97
|
+
}
|
98
|
+
```
|
99
|
+
|
100
|
+
### cluster
|
101
|
+
|
102
|
+
```json
|
103
|
+
{
|
104
|
+
"started_at": "2021-01-14T07:09:17Z",
|
105
|
+
"workers": 2,
|
106
|
+
"phase": 0,
|
107
|
+
"booted_workers": 2,
|
108
|
+
"old_workers": 0,
|
109
|
+
"worker_status": [
|
110
|
+
{
|
111
|
+
"started_at": "2021-01-14T07:09:24Z",
|
112
|
+
"pid": 64136,
|
113
|
+
"index": 0,
|
114
|
+
"phase": 0,
|
115
|
+
"booted": true,
|
116
|
+
"last_checkin": "2021-01-14T07:11:09Z",
|
117
|
+
"last_status": {
|
118
|
+
"backlog": 0,
|
119
|
+
"running": 5,
|
120
|
+
"pool_capacity": 5,
|
121
|
+
"max_threads": 5,
|
122
|
+
"requests_count": 2
|
123
|
+
}
|
124
|
+
},
|
125
|
+
{
|
126
|
+
"started_at": "2021-01-14T07:09:24Z",
|
127
|
+
"pid": 64137,
|
128
|
+
"index": 1,
|
129
|
+
"phase": 0,
|
130
|
+
"booted": true,
|
131
|
+
"last_checkin": "2021-01-14T07:11:09Z",
|
132
|
+
"last_status": {
|
133
|
+
"backlog": 0,
|
134
|
+
"running": 5,
|
135
|
+
"pool_capacity": 5,
|
136
|
+
"max_threads": 5,
|
137
|
+
"requests_count": 1
|
138
|
+
}
|
139
|
+
}
|
140
|
+
]
|
141
|
+
}
|
142
|
+
```
|
data/docs/systemd.md
CHANGED
@@ -1,21 +1,18 @@
|
|
1
1
|
# systemd
|
2
2
|
|
3
|
-
[systemd](https://www.freedesktop.org/wiki/Software/systemd/) is a
|
4
|
-
|
5
|
-
|
6
|
-
|
3
|
+
[systemd](https://www.freedesktop.org/wiki/Software/systemd/) is a commonly
|
4
|
+
available init system (PID 1) on many Linux distributions. It offers process
|
5
|
+
monitoring (including automatic restarts) and other useful features for running
|
6
|
+
Puma in production.
|
7
7
|
|
8
8
|
## Service Configuration
|
9
9
|
|
10
|
-
Below is a sample puma.service configuration file for systemd, which
|
11
|
-
|
12
|
-
|
10
|
+
Below is a sample puma.service configuration file for systemd, which can be
|
11
|
+
copied or symlinked to `/etc/systemd/system/puma.service`, or if desired, using
|
12
|
+
an application or instance-specific name.
|
13
13
|
|
14
|
-
Note that this uses the systemd preferred "simple" type where the
|
15
|
-
|
16
|
-
exit). See also, the
|
17
|
-
[Alternative Forking Configuration](#alternative-forking-configuration)
|
18
|
-
below.
|
14
|
+
Note that this uses the systemd preferred "simple" type where the start command
|
15
|
+
remains running in the foreground (does not fork and exit).
|
19
16
|
|
20
17
|
~~~~ ini
|
21
18
|
[Unit]
|
@@ -26,27 +23,39 @@ After=network.target
|
|
26
23
|
# Requires=puma.socket
|
27
24
|
|
28
25
|
[Service]
|
29
|
-
#
|
30
|
-
|
26
|
+
# Puma supports systemd's `Type=notify` and watchdog service
|
27
|
+
# monitoring, if the [sd_notify](https://github.com/agis/ruby-sdnotify) gem is installed,
|
28
|
+
# as of Puma 5.1 or later.
|
29
|
+
# On earlier versions of Puma or JRuby, change this to `Type=simple` and remove
|
30
|
+
# the `WatchdogSec` line.
|
31
|
+
Type=notify
|
32
|
+
|
33
|
+
# If your Puma process locks up, systemd's watchdog will restart it within seconds.
|
34
|
+
WatchdogSec=10
|
31
35
|
|
32
36
|
# Preferably configure a non-privileged user
|
33
37
|
# User=
|
34
38
|
|
35
|
-
# The path to
|
36
|
-
# Also replace the "<
|
37
|
-
|
39
|
+
# The path to your application code root directory.
|
40
|
+
# Also replace the "<YOUR_APP_PATH>" placeholders below with this path.
|
41
|
+
# Example /home/username/myapp
|
42
|
+
WorkingDirectory=<YOUR_APP_PATH>
|
38
43
|
|
39
44
|
# Helpful for debugging socket activation, etc.
|
40
45
|
# Environment=PUMA_DEBUG=1
|
41
46
|
|
42
|
-
#
|
43
|
-
#
|
44
|
-
#
|
45
|
-
ExecStart
|
47
|
+
# SystemD will not run puma even if it is in your path. You must specify
|
48
|
+
# an absolute URL to puma. For example /usr/local/bin/puma
|
49
|
+
# Alternatively, create a binstub with `bundle binstubs puma --path ./sbin` in the WorkingDirectory
|
50
|
+
ExecStart=/<FULLPATH>/bin/puma -C <YOUR_APP_PATH>/puma.rb
|
51
|
+
|
52
|
+
# Variant: Rails start.
|
53
|
+
# ExecStart=/<FULLPATH>/bin/puma -C <YOUR_APP_PATH>/config/puma.rb ../config.ru
|
46
54
|
|
47
|
-
# Variant: Use config file with `bind` directives instead:
|
48
|
-
# ExecStart=<WD>/sbin/puma -C config.rb
|
49
55
|
# Variant: Use `bundle exec --keep-file-descriptors puma` instead of binstub
|
56
|
+
# Variant: Specify directives inline.
|
57
|
+
# ExecStart=/<FULLPATH>/puma -b tcp://0.0.0.0:9292 -b ssl://0.0.0.0:9293?key=key.pem&cert=cert.pem
|
58
|
+
|
50
59
|
|
51
60
|
Restart=always
|
52
61
|
|
@@ -54,26 +63,31 @@ Restart=always
|
|
54
63
|
WantedBy=multi-user.target
|
55
64
|
~~~~
|
56
65
|
|
57
|
-
See
|
66
|
+
See
|
67
|
+
[systemd.exec](https://www.freedesktop.org/software/systemd/man/systemd.exec.html)
|
58
68
|
for additional details.
|
59
69
|
|
60
70
|
## Socket Activation
|
61
71
|
|
62
|
-
systemd and
|
63
|
-
|
64
|
-
|
65
|
-
|
66
|
-
|
67
|
-
|
72
|
+
systemd and Puma also support socket activation, where systemd opens the
|
73
|
+
listening socket(s) in advance and provides them to the Puma master process on
|
74
|
+
startup. Among other advantages, this keeps listening sockets open across puma
|
75
|
+
restarts and achieves graceful restarts, including when upgraded Puma, and is
|
76
|
+
compatible with both clustered mode and application preload.
|
77
|
+
|
78
|
+
**Note:** Any wrapper scripts which `exec`, or other indirections in `ExecStart`
|
79
|
+
may result in activated socket file descriptors being closed before reaching the
|
80
|
+
puma master process. For example, if using `bundle exec`, pass the
|
81
|
+
`--keep-file-descriptors` flag. `bundle exec` can be avoided by using a `puma`
|
82
|
+
executable generated by `bundle binstubs puma`. This is tracked in [#1499].
|
68
83
|
|
69
|
-
**Note:** Socket activation doesn't currently work on
|
70
|
-
|
84
|
+
**Note:** Socket activation doesn't currently work on JRuby. This is tracked in
|
85
|
+
[#1367].
|
71
86
|
|
72
|
-
|
73
|
-
|
74
|
-
|
75
|
-
|
76
|
-
above puma.service:
|
87
|
+
Configure one or more `ListenStream` sockets in a companion `*.socket` unit file
|
88
|
+
to use socket activation. Also, uncomment the associated `Requires` directive
|
89
|
+
for the socket unit in the service file (see above.) Here is a sample
|
90
|
+
puma.socket, matching the ports used in the above puma.service:
|
77
91
|
|
78
92
|
~~~~ ini
|
79
93
|
[Unit]
|
@@ -96,16 +110,42 @@ Backlog=1024
|
|
96
110
|
WantedBy=sockets.target
|
97
111
|
~~~~
|
98
112
|
|
99
|
-
See
|
113
|
+
See
|
114
|
+
[systemd.socket](https://www.freedesktop.org/software/systemd/man/systemd.socket.html)
|
100
115
|
for additional configuration details.
|
101
116
|
|
102
|
-
Note that the above configurations will work with Puma in either
|
103
|
-
|
117
|
+
Note that the above configurations will work with Puma in either single process
|
118
|
+
or cluster mode.
|
119
|
+
|
120
|
+
### Sockets and symlinks
|
121
|
+
|
122
|
+
When using releases folders, you should set the socket path using the shared
|
123
|
+
folder path (ex. `/srv/projet/shared/tmp/puma.sock`), not the release folder
|
124
|
+
path (`/srv/projet/releases/1234/tmp/puma.sock`).
|
125
|
+
|
126
|
+
Puma will detect the release path socket as different than the one provided by
|
127
|
+
systemd and attempt to bind it again, resulting in the exception `There is
|
128
|
+
already a server bound to:`.
|
129
|
+
|
130
|
+
### Binding
|
131
|
+
|
132
|
+
By default, you need to configure Puma to have binds matching with all
|
133
|
+
ListenStream statements. Any mismatched systemd ListenStreams will be closed by
|
134
|
+
Puma.
|
135
|
+
|
136
|
+
To automatically bind to all activated sockets, the option
|
137
|
+
`--bind-to-activated-sockets` can be used. This matches the config DSL
|
138
|
+
`bind_to_activated_sockets` statement. This will cause Puma to create a bind
|
139
|
+
automatically for any activated socket. When systemd socket activation is not
|
140
|
+
enabled, this option does nothing.
|
141
|
+
|
142
|
+
This also accepts an optional argument `only` (DSL: `'only'`) to discard any
|
143
|
+
binds that's not socket activated.
|
104
144
|
|
105
145
|
## Usage
|
106
146
|
|
107
|
-
Without socket activation, use `systemctl` as root (e
|
108
|
-
|
147
|
+
Without socket activation, use `systemctl` as root (i.e., via `sudo`) as with
|
148
|
+
other system services:
|
109
149
|
|
110
150
|
~~~~ sh
|
111
151
|
# After installing or making changes to puma.service
|
@@ -114,35 +154,35 @@ systemctl daemon-reload
|
|
114
154
|
# Enable so it starts on boot
|
115
155
|
systemctl enable puma.service
|
116
156
|
|
117
|
-
# Initial
|
157
|
+
# Initial startup.
|
118
158
|
systemctl start puma.service
|
119
159
|
|
120
160
|
# Check status
|
121
161
|
systemctl status puma.service
|
122
162
|
|
123
|
-
# A normal restart. Warning:
|
163
|
+
# A normal restart. Warning: listener's sockets will be closed
|
124
164
|
# while a new puma process initializes.
|
125
165
|
systemctl restart puma.service
|
126
166
|
~~~~
|
127
167
|
|
128
|
-
With socket activation, several but not all of these commands should
|
129
|
-
|
168
|
+
With socket activation, several but not all of these commands should be run for
|
169
|
+
both socket and service:
|
130
170
|
|
131
171
|
~~~~ sh
|
132
172
|
# After installing or making changes to either puma.socket or
|
133
173
|
# puma.service.
|
134
174
|
systemctl daemon-reload
|
135
175
|
|
136
|
-
# Enable both socket and service so they start on boot. Alternatively
|
137
|
-
# you could leave puma.service disabled and systemd will start it on
|
138
|
-
# first use (with startup lag on first request)
|
176
|
+
# Enable both socket and service, so they start on boot. Alternatively
|
177
|
+
# you could leave puma.service disabled, and systemd will start it on
|
178
|
+
# the first use (with startup lag on the first request)
|
139
179
|
systemctl enable puma.socket puma.service
|
140
180
|
|
141
|
-
# Initial
|
181
|
+
# Initial startup. The Requires directive (see above) ensures the
|
142
182
|
# socket is started before the service.
|
143
183
|
systemctl start puma.socket puma.service
|
144
184
|
|
145
|
-
# Check status of both socket and service.
|
185
|
+
# Check the status of both socket and service.
|
146
186
|
systemctl status puma.socket puma.service
|
147
187
|
|
148
188
|
# A "hot" restart, with systemd keeping puma.socket listening and
|
@@ -155,8 +195,8 @@ systemctl restart puma.service
|
|
155
195
|
systemctl restart puma.socket puma.service
|
156
196
|
~~~~
|
157
197
|
|
158
|
-
Here is sample output from `systemctl status` with both service and
|
159
|
-
|
198
|
+
Here is sample output from `systemctl status` with both service and socket
|
199
|
+
running:
|
160
200
|
|
161
201
|
~~~~
|
162
202
|
● puma.socket - Puma HTTP Server Accept Sockets
|
@@ -187,70 +227,14 @@ Apr 07 08:40:19 hx puma[28320]: * Activated ssl://0.0.0.0:9234?key=key.pem&cert=
|
|
187
227
|
Apr 07 08:40:19 hx puma[28320]: Use Ctrl-C to stop
|
188
228
|
~~~~
|
189
229
|
|
190
|
-
## Alternative Forking Configuration
|
191
|
-
|
192
|
-
Other systems/tools might expect or need puma to be run as a
|
193
|
-
"traditional" forking server, for example so that the `pumactl`
|
194
|
-
command can be used directly and outside of systemd for
|
195
|
-
stop/start/restart. This use case is incompatible with systemd socket
|
196
|
-
activation, so it should not be configured. Below is an alternative
|
197
|
-
puma.service config sample, using `Type=forking` and the `--daemon`
|
198
|
-
flag in `ExecStart`. Here systemd is playing a role more equivalent to
|
199
|
-
SysV init.d, where it is responsible for starting Puma on boot
|
200
|
-
(multi-user.target) and stopping it on shutdown, but is not performing
|
201
|
-
continuous restarts. Therefore running Puma in cluster mode, where the
|
202
|
-
master can restart workers, is highly recommended. See the systemd
|
203
|
-
[Restart] directive for details.
|
204
|
-
|
205
|
-
~~~~ ini
|
206
|
-
[Unit]
|
207
|
-
Description=Puma HTTP Forking Server
|
208
|
-
After=network.target
|
209
|
-
|
210
|
-
[Service]
|
211
|
-
# Background process configuration (use with --daemon in ExecStart)
|
212
|
-
Type=forking
|
213
|
-
|
214
|
-
# Preferably configure a non-privileged user
|
215
|
-
# User=
|
216
|
-
|
217
|
-
# The path to the puma application root
|
218
|
-
# Also replace the "<WD>" place holders below with this path.
|
219
|
-
WorkingDirectory=
|
220
|
-
|
221
|
-
# The command to start Puma
|
222
|
-
# (replace "<WD>" below)
|
223
|
-
ExecStart=bundle exec puma -C <WD>/shared/puma.rb --daemon
|
224
|
-
|
225
|
-
# The command to stop Puma
|
226
|
-
# (replace "<WD>" below)
|
227
|
-
ExecStop=bundle exec pumactl -S <WD>/shared/tmp/pids/puma.state stop
|
228
|
-
|
229
|
-
# Path to PID file so that systemd knows which is the master process
|
230
|
-
PIDFile=<WD>/shared/tmp/pids/puma.pid
|
231
|
-
|
232
|
-
# Should systemd restart puma?
|
233
|
-
# Use "no" (the default) to ensure no interference when using
|
234
|
-
# stop/start/restart via `pumactl`. The "on-failure" setting might
|
235
|
-
# work better for this purpose, but you must test it.
|
236
|
-
# Use "always" if only `systemctl` is used for start/stop/restart, and
|
237
|
-
# reconsider if you actually need the forking config.
|
238
|
-
Restart=no
|
239
|
-
|
240
|
-
[Install]
|
241
|
-
WantedBy=multi-user.target
|
242
|
-
~~~~
|
243
|
-
|
244
230
|
### capistrano3-puma
|
245
231
|
|
246
|
-
By default,
|
247
|
-
|
248
|
-
|
249
|
-
|
250
|
-
|
251
|
-
|
252
|
-
also that the configured `User` should likely be the same as the
|
253
|
-
capistrano3-puma `:puma_user` option.
|
232
|
+
By default, [capistrano3-puma](https://github.com/seuros/capistrano-puma) uses
|
233
|
+
`pumactl` for deployment restarts outside of systemd. To learn the exact
|
234
|
+
commands that this tool would use for `ExecStart` and `ExecStop`, use the
|
235
|
+
following `cap` commands in dry-run mode, and update from the above forking
|
236
|
+
service configuration accordingly. Note also that the configured `User` should
|
237
|
+
likely be the same as the capistrano3-puma `:puma_user` option.
|
254
238
|
|
255
239
|
~~~~ sh
|
256
240
|
stage=production # or different stage, as needed
|
@@ -260,3 +244,4 @@ cap $stage puma:stop --dry-run
|
|
260
244
|
|
261
245
|
[Restart]: https://www.freedesktop.org/software/systemd/man/systemd.service.html#Restart=
|
262
246
|
[#1367]: https://github.com/puma/puma/issues/1367
|
247
|
+
[#1499]: https://github.com/puma/puma/issues/1499
|