RoadRunner
-
|
When I use memory as the jobs driver, I have config: jobs:
consume: ["queue1", "queue2"]
pool:
num_workers: 1
pipelines:
queue1:
driver: memory
config:
priority: 10
prefetch: 1
queue2:
driver: memory
config:
priority: 10
prefetch: 1When I push 2 messages to queue1, I find that there is only one php worker.php process, which is ok. But When I use nats as the jobs driver, I have config: jobs:
consume: ["queue1", "queue2"]
pool:
num_workers: 1
pipelines:
queue1:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
stream: default
queue2:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
stream: defaultI find that there are two php worker.php process. Unless I configure different streams for different pipelines. I don't really understand how the pool worker works, but I think it should be explicitly written in the documentation or restricted in the code to not allow the same stream to be used |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 19 replies
-
|
Hey @shellphy 👋 |
Beta Was this translation helpful? Give feedback.
-
|
ps aux |
Beta Was this translation helpful? Give feedback.
-
|
I know that in the nats jobs driver, each different pipeline should have a different stream name |
Beta Was this translation helpful? Give feedback.
-
|
Are you sure that you stopped all RR instances? And have only 1. Driver does not have workers, only JOBS plugin allocates them. So, there is no difference what driver you'd use, memory, nats, etc. |
Beta Was this translation helpful? Give feedback.
-
I've double-checked, please, be sure that you don't have any other RR instances.
|
Beta Was this translation helpful? Give feedback.
-
|
Again, how this related to the |
Beta Was this translation helpful? Give feedback.
-
|
Please move your messages to the correct thread. Then, create a ticket, describe your problem, and if you're willing to fix that problem (👍) create attach/create a PR. |
Beta Was this translation helpful? Give feedback.
-
|
config is: nats:
addr: nats://nats:4222
jobs:
consume: ["queue1", "queue2"]
num_workers: 1
pipelines:
queue1:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queue
queue2:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queuesend If I configure different streams for different pipelines, it will be normal: ...
queue1:
...
stream: default_queue1
...
queue2:
...
stream: default_queue2
...So I think we should explain this situation in the documentation, with special attention. |
Beta Was this translation helpful? Give feedback.
-
|
There is no connection between the number of workers and the stream name in the NATS plugin. I've double-checked that case, shared a screenshot with you above. In your case, this is either a misconfiguration or 2 RR instances running (might be one of it in the background). |
Beta Was this translation helpful? Give feedback.
-
|
Moreover, you have an invalid configuration: https://github.com/roadrunner-server/roadrunner/blob/master/.rr.yaml#L1236, num_workers should be under the pool section. |
Beta Was this translation helpful? Give feedback.
-
|
I ran the test again If I configured debug: jobs
pool:
num_workers: 1
debug: trueThere will be two worker process. but if I set debug=false, it will be fine. |
Beta Was this translation helpful? Give feedback.
-
|
config is: version: '3'
rpc:
listen: 'tcp://127.0.0.1:6001'
server:
command: 'php worker.php'
nats:
addr: nats://127.0.0.1:4222
logs:
mode: development
level: debug
jobs:
consume: ["queue1", "queue2"]
pool:
num_workers: 1
debug: true
pipelines:
queue1:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queue
queue2:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queueproducer.php: <?php
use Spiral\Goridge\RPC\RPC;
use Spiral\RoadRunner\Jobs\Jobs;
require "vendor/autoload.php";
$jobs = new Jobs(
RPC::create('tcp://127.0.0.1:6001')
);
$queue = $jobs->connect('queue1');
$task = $queue->create("dummy", 'hello');
$result = $queue->dispatch($task);
var_dump($result);worker.php <?php
use Spiral\Goridge;
ini_set('display_errors', 'stderr');
require "vendor/autoload.php";
$consumer = new Spiral\RoadRunner\Jobs\Consumer();
while ($task = $consumer->waitTask()) {
try {
echo "received " . $task->getPayload();
sleep(20);
} catch (\Throwable $e) {
var_dump($e->getMessage());
}
} |
Beta Was this translation helpful? Give feedback.
-
|
.rr.yaml: version: '3'
rpc:
listen: 'tcp://127.0.0.1:6001'
server:
command: 'php worker.php'
nats:
addr: nats://127.0.0.1:4222
logs:
mode: development
level: debug
jobs:
consume: ["queue1", "queue2"]
pool:
num_workers: 2
debug: false
pipelines:
queue1:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queue
queue2:
driver: nats
config:
priority: 10
prefetch: 1
delete_after_ack: true
subject: default_queue
stream: default_queueproducer.php: <?php
use Spiral\Goridge\RPC\RPC;
use Spiral\RoadRunner\Jobs\Jobs;
require "vendor/autoload.php";
$jobs = new Jobs(
RPC::create('tcp://127.0.0.1:6001')
);
$queue = $jobs->connect('queue1');
$task = $queue->create("dummy", 'hello');
$result = $queue->dispatch($task);
var_dump($result);worker.php: <?php
use Spiral\Goridge;
ini_set('display_errors', 'stderr');
require "vendor/autoload.php";
$consumer = new Spiral\RoadRunner\Jobs\Consumer();
while ($task = $consumer->waitTask()) {
try {
echo "received " . $task->getPayload();
sleep(20);
} catch (\Throwable $e) {
var_dump($e->getMessage());
}
}test: log: You can see two messages: |
Beta Was this translation helpful? Give feedback.
-
|
2 messages - 2 workers. |
Beta Was this translation helpful? Give feedback.
-
|
But I send only one message to queue1 |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
I see, this is how nats work, but I think it should be emphasized in the roadrunner documentation. |
Beta Was this translation helpful? Give feedback.
- There is no bug:
I've double-checked, please, be sure that you don't have any other RR instances.Purge.