I am trying to eliminate this idea as a cause of a problem.
About a week and a half ago, we transferred all files to a new virtual server. No problems. About a week ago, we updated from the 22.6001 build to the 22.6021 build. No problems.
Last Thursday morning, we introduced our first API plug-in to a subset of 6 users in our studio of 42 users. This plug-in adds a company menu item with several tools to the ArchiCAD menu bar. Some of these tools are event driven reactors that use:
To provide this Add-On, I changed the [Options > Add-On Manager > Use designated Add-Ons folder] to be a network location. Friday morning, some users started having trouble in their Teamwork projects. The first user [regardless of being part of the subset of 6] could open the model. Any other user after that [regardless of being part of the subset of 6] could not get in to the same model.
We remade the hot links that were made with the 6001 build with the 6021 build. No help.
This morning, we excluded the API Add-On from the subset of 6 by
[Add_on Manager > uncheck the plug-in > click Remove].
After removing the Add-On, all users are able to get in the teamwork model, but still experience crashing. Crashing typically occurs when relinking hot links.
We have had other users that are not part of the subset of 6 that are not crashing, but get an error message that states "Your computer does not have enough memory to perform the operation." when they join the teamwork project. Memory is pegged at 94% even though that machine has 64GB.
- - Are there sharing violations I'm not aware of when using a single network location for API storage and access?
- - Could this event driven, model monitoring Add-On be showing symptoms of memory leaks? Crashing and memory problems occur prior to any use of the menu. Will began memory leak detection now.
- - Could this possibly be related to the Add-On at all?