Large Folder Count Under The FIN Root Crashes PowerGEOPAK / Workspace

We have a FIN in which the total folder count is over 2200 folders. This is causing PGPK to crash across all users on opening any .dgn file under the FIN. The crashing is 100% consistent and reproducible. It is not any specific folder - reducing the overall folder count by removing folders from anywhere in the folder tree will stop the crashing. The *file* count seems to be immaterial - only the folder count seems to be an issue. However, I suspect that there could be a file count limit also since the problem very likely involves a memory / buffer size limitation.

It does not crash outside of the FDOT Workspace - presumably because of the file/folder intercept scanning the Workspace does.

The count threshold does not appear to be necessarily a specific number, nor linear - ie. it is not necessarily the case that while adding a single folder would break PGPK, removing a different single folder would restore PGPK. While we got to a point where we could make/break PGPK by the add/remove of a single empty folder at the top level right under the FIN, we had to remove multiple folders if we removed from a deeper folder level. In other words, we could add one top-level folder and break PGPK, then remove a different single top-level folder and PGPK would work again. But removing a single folder from a deeper level would not restore PGPK - we had to remove a chunk of folders to get it restored. Thus I concluded that it is some function / permutation of width x depth rather than a rigid count. I did not fully flesh out the permutation as it would be a very tedious endeavor.

We got into this state by running ICPR calcs in a folder under the drainage folder. ICPR can generate a large number of folders when calculating a larger model.

Obviously, this happened while grinding out a 90% submittal over this past weekend that was due Monday.  (Is it true that you guys have to enter your submittal dates when initially setting up a job so that Bentley / FDOT knows when to crash the software, corrupt/delete files, etc, at the worst possible time?   (c:  )   There were 4 or 5 people in the job working at the same time, one of them running ICPR calcs. When the calcs tipped the folder count threshold, everyone's PGPK crashed on the next file open, and production came to a halt. Other jobs worked fine. Only this job was affected.

After some digging that pointed to something in the job was crashing it, I moved out to a duplicate FIN folder, with a _meta_info folder included, only the basic folders needed to get the job back in to production. Once production was finished, I researched/resolved the problem and told the guys they had to trim out some fodders, maybe by zipping some stuff, before packaging the job for submittal.

This is too esoteric of a problem to expect a fix, and too easy to avoid / work around. So I do not post this with the expectation of a fix, but rather to inform others of the potential trap.

Mike Champion

Orlando