php-enqueue/Lobby

Paweł Niedzielski

@steveb_gitlab

if you're using docker then you already have a supervisor process (docker itself)

TheJenne18

@TheJenne18

And in both cases if a job takes longer than 5 minutes another process will run the same job. I would think it is changeble in the config or via a cli parameter

Paweł Niedzielski

@steveb_gitlab

oh ok, now I get it

unfortunately I'm not really familiar with job queue package and how it locks jobs internally

but it definitely looks like it's missing a job timeout setting

TheJenne18

@TheJenne18

https://github.com/php-enqueue/redis/blob/master/RedisConnectionFactory.php#L41-L42

Alexander Trauzzi

@atrauzzi

Hey @makasim! :wave: :smile: -- I was wondering, is it possible to disable any attempt to use delay when using the Laravel enqueue package? I'm running into an issue where my failed jobs are trying to use the delay functionality over the stomp driver. I don't particularly need there to be a delay, so having an option to turn it off altogether would probably be fine...

Alexander Trauzzi

@atrauzzi

Ahh, think I might have tracked it down.

Alex

@atrigub

Hi guys, Could you explain me how receive_timeout should work? Why it works even when we have a lot messages in queue

Niv Penso

@nivpenso

Hi guys, I'm curious to know how does the release from enqueue-dev to the production packages work?

Paweł Niedzielski

@steveb_gitlab

@nivpenso see https://github.com/php-enqueue/enqueue-dev/blob/master/bin/subtree-split and https://github.com/splitsh/lite

Niv Penso

@nivpenso

@steveb_gitlab cool thank you for the info.. is this script running automatically or is it being executed manually?

This message was deleted

Max Kotliar

@makasim

It is executed manually

Niv Penso

@nivpenso

ok got it :)

thanks

Luciano Carvalho

@lucianoacjunior_twitter

Hey folks!

Regarding the transport of using rdkafka (library in C), how can we know up to which version enqueue/rdkafka supports?

Niv Penso

@nivpenso

https://packagist.org/packages/enqueue/rdkafka

I assume that you can check it in the requires section under ext-rdkafka

Luciano Carvalho

@lucianoacjunior_twitter

Excellent, thank you very much

Niv Penso

@nivpenso

Hi, is there any forecast when is the next release?

Paweł Niedzielski

@steveb_gitlab

@nivpenso Max pushed a release a few hours ago.

Niv Penso

@nivpenso

Thanks, I just integrated the new version. Well Done

Édipo Costa Rebouças

@edipoReboucas

Hi guys, can someone send me some docs on how to use extensions (like Doctrine) with SimpleClient without symfony bundle? I'm almost copying / pasting the Simple Client and making customizations.

Édipo Costa Rebouças

@edipoReboucas

For now I created a QueueConsume wrapper that pass additional extensions to consume method of QueueConsume wrapped,

Jan Jouke Tjalsma

@janjouketjalsma

Hi, the current enqueue kafka transport requires librdkafka >= 1.0.0. Is there a version of the transport that is safe to use with a lower version of librdkafka (like 0.85 or 0.11)?

Max Kotliar

@makasim

The kafka was upgraded in the latest release (0.10.4). So any prior should do.

Jan Jouke Tjalsma

@janjouketjalsma

Thanks! Tried that and didn't see issues, but it's nice to be sure :)

Chris

@Kalyse

I'm not sure if I have fundamental misunderstanding of enqueue, I want to use the async event listeners with Symfony. I have events which currently listen to new entity creations. Like "New Order" etc. So, I would normally dispatch the events on the new entities and pass through the unflushed entity, but then the subscriber which enqueue get's obviously get's the unpersisted entity tooo.. it seems there is no way to use async in the same way

Niv Penso

@nivpenso

I don't think that this is relevant specifically to enqueue

Chris

@Kalyse

Yeah, I had a step back, and considered, that I jsut need to stop sending my entities as messages, to avoid the unpersisted entity problem

And to use a transformer on my Events before hand

Niv Penso

@nivpenso

if you are using a different object of EntityManagerInterface (e.g subscriber in a worker on different process) you must persist your related entities as well before flushing it using the entity manager

Well, this is your architectural decision of whether to send entities or not. either way, you would need to call persist in order to use in front of the DB

Chris

@Kalyse

I have a Transformer'enqueue.event_transformer'

It throws an exception... How to I catch it from my consume so I can reject the message?

I've basically got an unrecoverable message in the queue that I can't kick out.

If my transformer throws a catachable exception, I just want it rejected

Actually, this is relevant, the consume is handled by https://php-enqueue.github.io/bundle/async_events/ it's an async processor

Chris

@Kalyse

Solved it. I ended up needing to catch the relevant Exception in my Transformer, then throwing a different type, and registeriung a extension which implements ProcessorExceptionExtensionInterface and returns a Result::Reject()

Valentin Saik

@svbackend

Hello @makasim could you please have a look to php-enqueue/enqueue-dev#1121 ? would be really cool if you can help with questions mentioned there, as I'm kind of blocked because of forma-pro docker image that doesn't support php8 yet (https://github.com/formapro/docker-nginx-php-fpm), I'm not sure you're still working there so should we rely on this docker image at all? Or would be better to move docker images to your (or enqueue) repository?

Adnan Siddiq

@adnansiddiq

Hello, Can I use php-enqueue/laravel-queue in Laravel 5.2 with php 7.0

Andrew

@andrewmy

Hey @makasim, I've sent a PR php-enqueue/enqueue-dev#1126
Your travis is still waiting but has run through on my side https://travis-ci.com/github/andrewmy/enqueue-dev/builds/213201382
It appears that the only failing tests are for pubsub, I guess nothing to do with my changes 🤷‍♂️

1 reply

Max Kotliar

@makasim

travis-ci.org is known to be slow. As far as I know it is kind of legacy and all project should migrate to travis-ci.com

Unfortunately I have no time to do the migration in near future.

There is also an issue with docker hub images. Because of it's new polices it is useless for OS and breaks CI atm.

That should be addressed too.

Without reliable CI I cannot merge PRs.

Andrew

@andrewmy

Ouch. Heard about the travis crap, probably best to move to gh actions

Max Kotliar

@makasim

Good idea. Though It's hell of a lot work.

Andrew

@andrewmy

There's a bug in php-rdkafka, reported in their gitter — the message headers are deserialized to all strings when they can also be ints and nulls. E.g.:

RdKafka\Message {#485
  +err: 0
  +topic_name: "messages"
  +timestamp: 1610657820673
  +partition: 0
  +payload: "{"body":"some_body","properties":[],"headers":{"Content-Type":"application\/json","name":"some_name","id":"8d2977e3-695f-45e9-a09f-f1541db273da","timestamp":1610657820,"retry_count":null}}"
  +len: 188
  +key: null
  +offset: 2
  +headers: array:5 [
    "Content-Type" => "application/json"
    "name" => "some_name"
    "id" => "8d2977e3-695f-45e9-a09f-f1541db273da"
    "timestamp" => "1610657820"
    "retry_count" => ""
  ]
  errstr: "Success"
}

See timestamp and retry_count.

Now the enqueue consumer prefers those headers to its own deserialization result. Would it make sense to drop the header merge? Or reverse the preference?

Where communities thrive

People

Repo info

Activity