Starting from last Friday, my local Docker env keeps throwing 502 Bad Gateway errors when trying to open Yves URLs. It might be related to the latest Docker update. Anybody else experiencing this issue?

Same here. Two different projects that show this issue.

We can confirm. On our side, this seems to be related to an update of the spryker/php docker images, which got a new release on friday. Looks like we cannot easily roll back that update, as I cannot find old digest-ids of the images from before friday

Same here. Seems to be memory issues. This is what I get from the logs

[yves] [yves_eu] - zend_mm_heap corrupted
[yves] [yves_eu] - [11-Oct-2021 09:05:26] WARNING: [pool worker] child 41 exited with code 1 after 58.771321 seconds from start
[yves] [yves_eu] - [11-Oct-2021 09:05:26] NOTICE: [pool worker] child 45 started

Do you know if opcache is enabled for you? I found something weird earlier this year: https://github.com/phpstan/phpstan/issues/4881#issuecomment-824068414

It's the default Spryker configuration. OpCache is enabled but the

validate_timestamps

and

revalidate_frequency

options are set to render the OpCache inactive

Yup I get the same error:

zed-backoffice] [backoffice_eu] - zend_mm_heap corrupted
[proxy] [frontend] - 2021/10/11 09:40:49 [error] 9#9: *37 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 172.25.0.20, server: backoffice.de.henry.local, request: "GET /security-gui/login HTTP/1.1", upstream: "<fastcgi://172.25.0.17:9001>", host: "backoffice.de.henry.local", referrer: "<http://backoffice.de.henry.local/security-gui/login>"
[zed-backoffice] [backoffice_eu] - [11-Oct-2021 09:40:49] WARNING: [pool debugger] child 17 exited with code 1 after 7872.539548 seconds from start
[proxy] [gateway] - 172.25.0.1 - - [11/Oct/2021:09:40:49 +0000] "GET /security-gui/login HTTP/1.1" 502 922 "<http://backoffice.de.henry.local/security-gui/login>" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.107 Safari/537.36" "-"
[zed-backoffice] [backoffice_eu] - [11-Oct-2021 09:40:49] NOTICE: [pool debugger] child 19 started

Hi @kind-vegetable-16151 @adventurous-raincoat-66045 we faced with similar problems on our project and according to latest changes for docker-images blackfire were affected. So in our case disabling of blackfire extension was helpful and solved the issue with zend_mm_heap.

Hey @wonderful-rainbow-33332, thanks for the hint

Trying that one out currently

@kind-vegetable-16151 please see: https://sprykercommunity.slack.com/archives/CH2HA71C5/p1633950070134600?thread_ts=1633941069.132000&cid=CH2HA71C5

Nice

Yes, gentlemen, it is confirmed that the problem is connected to recently upgraded blackfire extension. The issue’s nature is sporadical that could be a reason why it was not caught on the development stage before releasing. The solution is already in place: the commit is reverted, the images are rebuilt with the previous state.

docker/sdk pull && docker/sdk up

should do the trick by updating the base image to the latest one. Please, also if you have docker caches in your CI/CD make sure your pipelines are using proper images.

Nice, thanks @high-pencil-62400

Thank you!

Great, thanks a ton!

Please, update everybody here if the solution solves the problem, Thanks, gentlemen.

Works like a charm 👍

Yep.

Yes

Awesome! Keep be so communicative as you are today. Good job on catching this in time.

Fixed our CI pipeline, too

👍