They better figure it out before Flex2 date, or I’ll basically be down. We are highly automated and dependent on functions.
It is supposed to spread the changes to other pods on the next server call.
PRB0288709 describes a problem I have ran into intermittently before, where the changes aren’t seen on the server, which in the past has been resolved by restarting.
I feel this problem is different (although maybe related), because in this case nothing is changing. The functions are there. The application just suddenly can’t find them.
This is affecting flex 2 systems that are still on 2025.2 today so its not related to 2026.100.
Shit…. I should check my logs…
The issue seems to be increasing though, I wasn’t having this issue before
Well that’s the fun part. It doesn’t show up in the server logs. Because the application can’t reach the server. That’s why its coming back with 404. Either that or there is a server log error, but the logs from that server aren’t getting posted to server file download.
Same.
I meant my own. If some work and some don’t, I should pick up when a working one calls one that doesn’t.
My curiosity is peaking on whether yhis is an architecture or infrastructure problem.
or…
Currently having an issue with a function not running and checked the logs and was getting inconsistent results with 404s and one time it went through. I wonder if this is related.
Here’s the summary of my ticket I submitted which has yet to get assigned to anyone 
We have a custom Epicor Function (Library: [Libray Name], Function: [Function Name] ) called from a UD Screen (“Update Orders…” button). After identifying and fixing an EF Core compatibility bug in the function’s code (unmaterialized LINQ query causing “already an open DataReader” errors — a known 2026.100 EF Core migration issue), we republished and repromoted the Function Library for both authorized companies. Since then, identical calls to the same endpoint are returning inconsistent results: some succeed cleanly, some return 404 Not Found, and some still throw the original pre-fix exception — suggesting not all app server replicas are running the current compiled version of the Function.
Yep, sounds like the same issue.
TP on a roll. Plastic seats. Bright clean pine siding. Hi falutin’!
I updated the title and I suggest all you having this issue that have created tickets report that this is not an isolated incident and send them a link to this thread.
This looks like a major problem that needs to be addressed quickly.
I am not currently experiencing it, but if I was, it would be costing me real money.
It would be great to include exact time when you see this, so logs can be reviewed at correct point of time
Nothing shows up in the logs.
I want to check http logs to see how the call is routed. If it is really just a balanced issue
Thanks for being so active and interested in here @Olga