Dashboards from test are crossing over to prod

Has anyone had dashboard assemblies that were made in a (Pilot) system open up in production? I don’t know what details would be helpful to share, but this is going to be a problem. I’m working on some development in pilot, and and working on translating it to production, but somehow the deployed dashboard is pulling the pilot one instead of the test one. It’s grabbing production data, but the format is from test. I don’t know what to say to support if I were to call them either.

Has anyone seen this before?

So after some digging around I found where the DLL was being stored. It was in the file location C:\ProgramData\Epicor\epicor10-1-808\3.1.600.0\RAPAT\shared. I deleted the DLL in question and now it pulls up the right one.

It looks like Pilot and Production are using the same files.?.?..

I’m thinking we might have something installed goofy on our system. I would think that the cache locations for each database should be separate, does that sound right? For some reason for us it is not.

One workaround we had to do was to put a skip on the Pilot side for looking at the config file. Otherwise, anytime we opened up production, then pilot, then back again it would have to pull an update. Maybe this is because the system is trying to share something it shouldn’t?

Any ideas would be appreciated.

Not sure where the problem lies but can confirm it happens.

I have a similar issue at a site with Live, Test and Pilot.
If I deploy a revised dashboard in Pilot, it impacts Live and vice versa.

I’m guessing a step was missed the last time Pilot was refreshed.
For now I just remember not to deploy dashboard revisions in Pilot.
and I’m hoping this will be fixed automatically the next time Pilot is refreshed,
Since the local admin has refreshed Test recently and deploying dashboards are not an issue there.

Are you seeing it on others computers (those that don’t use pilot) as well as yours? Or only on yours? It seems to be related to the local cache so if someone never opens the one from Pilot on their computer, I don’t think (I hope anyways) it will be a problem for anyone but the developer. But I’m not sure. Just trying to figure out how the system works.

It impacts all workstations… I assume it’s a config file on the server didn’t get edited during the last refresh but… I don’t have admin access at this site.

As far as I know the local cache is just where the latest copy from the server gets stored.

What is it in the config file that needs to get edited? I’m not sure if we do any edit in the config file. In fact, we put a /skip in the shortcut to not look at the pilot config file because it causes the system to reload things every time you switch from one to the other.

not “THE config file” but “A config file” intended as a generic term

Hopefully someone else can supply specifics.

I don’t have admin rights at this site.
Right now I can only guess it’s something in the System Agent paths, Task Agent or Application Pool.

1 Like

I don’t think you’re alone in this, by default the local cache will be in the same place for all environments. The reason I think your test system refresh doesn’t affect it will be because you’ve not gone into dashboard maintenance and re-deployed all of the compiled dashboards. If you do that, you’ll see the test ones not the live ones.

Clear Client Cache from within Epicor should resolve it, as next time you open the dashboard in Live it will download the .dll again.

In terms of fixing moving forwards, i think in the local .sysconfig file on the client PC you can specify a temp folder to use. Set that folder to be different for each environment.

This is one of the problems that we are having is that we will make a change and redeploy, then clear the cache and the new one shows up the first time, but any time after that, it goes back to the old one. So I think the cache is only clearing the dashboards from the system that you are in, but it’s reading all of them. Maybe I should try clearing the cache in both sytems and see if it sticks? That’s maybe why it’s doing what it’s doing. (then I can tell support what the problem is because I have a call in on this, they recreated the issue, but I don’t think they have a fix for it)

I was told that you shouldnt have PILOT/TEST/DEMO on the same server as your Prod DB Server, for this reason of common files.

I asked the question early this year.

1 Like

It’s not the server that’s the issue, its the local cache on the client.

Look in the config editor on a client - there is a field for AlternateClientCache. Try setting this for each environment on 1 test client - then fully clear the cache. Then open Epicor for each environment and check if it saves temp files to the specified directories.

1 Like

It does change where it saves! Thanks, I’ll do some testing and see if it fixes my problems.