Enfocus Community

Hi,

Every day, I experience Socket Hangup errors in Switch 2024 Fall. This occurs primarily, but not exclusively, in the HTTP request module (version 20.3).

However, the issue does not seem to lie in the request itself but rather in writing metadata after the request has been completed. I base this on the fact that when I send an order to an MIS system via a POST request, it is successfully received by the MIS system, and the messages in Switch confirm this as well.

I've experimented with other versions of the HTTP module, but that didn't help.

Enfocus is aware of the issue and has given it the highest priority, but a solution is expected to be included in the next Switch release in June/July.

I’m curious to know who else is experiencing this issue and whether anyone has found a workaround.

@RobertHumphrey: If you can afford to lose some functionality, you can downgrade to version 15. That version seems to work fine. Read the docs carefully to know what you will be missing.

Hi old colleague

I’ve seen these socket errors in all sorts of variations and across multiple Switch 2024 Fall servers, so at least you are not alone

Most of the time it’s HTTP request (20.3), but indeed not exclusively;
Unknown error Creating job error: socket hang up
Unknown error Cannot change dataset for child job: socket hang up
As of today, 'Wait for Asset' joined the party;
Creating job error: socket hang up Error: Creating job error: socket hang up at _0x3d6d0e._createJob
I haven't found a workaround, after a while I gave up and routed the error jobs out of Problem jobs to 'auto retry' these.

Hi Sander,

Good to know that I'm not alone on this one... But I already knew

Version 15 runs stable, but it doesn't have all the features like 20.3.

Since the request (with 20.3) usually succeeds, it is not always an option to retry, since that may result in jobs being sent twice to production etc.

Robert