I am at my wits end with this one. This has been occurring for select users for a couple months.

We have two locations running the same file. Each with it's own copy of the file on it's own server. This is happening a couple times a day for one out of 20 users at one location and once in a great while for one user out of 12 at the other location.

I have a layout with a scripted find over multiple tables. The user enters the find criteria into a global field and fires the script by clicking a button or hitting enter. When the search executes it appears to find the record being searched for and then FileMaker freezes up and closes without any kind of dialog. When the database is reopened the record in question is gone.

Initially the script was using quick find to do the search. I read some threads that mentioned quick find crashing and closing filemaker so I thought it might be the culprit. I rewrote the find script using multiple find requests instead. Even with the new completely rewritten script it continues to happen. In all we have over 30 users across two locations that use this layout all day everyday and yet it is only happening to one person regularly. The other instances have been isolated and maybe occurred like once a month.

The script has been completely rewritten and in both instances of the script there is absolutely no delete command anywhere.

The user in question does not have permission to delete records in that table at all

The result when it does happen is always that FileMaker locks up and crashes, apparently obliterating the current record in the process.

For the user who is experiencing it frequently we even tried switching out his workstation computer but it still continues to happen to him.

 

I keep scaling back the find to search fewer and fewer fields but that is certainly not a solution.

Both locations are running Filemaker Server 17 on windows Server 16. The clients are all Filemaker 17 on Windows 10.

 

Even looking at my scripting and everything involved the fact that it is resulting in Filemaker crashing and the record being deleted is crazy. I'm stumped...

 

 

You've had that user switch to a different workstation and it still crashes?

Yes

Um. Anything specific to the record that results? For example, does that user always look for a record others do not? (btw, I haven't had issues with QF).

Font corruption was my guess, but if it happens on a different workstation...

Recovering a copy report any corruption?

In a way yes. The table the layout is based on is a join table for sales leads. The join table joins the lead record to a set record. The leads are assigned to sets in groups of 100 and those sets are then assigned to a sales person. When assigning the set it actually puts the users ID in the set record and loops through all the lead assignment records and puts the user ID in each one. So the records are all assigned to a user. The find script runs the search and up until earlier today it would produce a found set and then do another find request to constrain the found set to only those with the user 's ID. Earlier today I restructured the script again to include the users ID in each record request of the find to begin with. Thus eliminating the need for a constrain. I had wanted to have the find constrain conditionally based on the users permission levels. In other words a manager would get all results, where a sales person would only get their own records. Either way I don't know if eliminating the constrain find will help but I'm trying anything I can. Now I have to wait for the user to report if it happens again.

In the end though I can't imagine that the user ID could have anything to do with things but who knows at this point.

 

If the RLA is defined that a record is only viewable to the user if the gUserId = recordUserId, then you don't need the constrain.

Well, let's see what removing the constrain does.

The user in question has knocked off for the day, so I guess we'll see what happens tomorrow.

Thanks

I got away from this for a while but It still remains unresolved. After rewriting the find script again to remove the constrain found set, there was no change. The situation still remains 2 locations running copies of the same solution with different data. Each location has one person experiencing the issue out of 12 or more that use the system for the same job working from the same layout.

For the 2 users in question, when they run the find script it seems to find the record in question and then filemaker crashes and when it is reopened the record in question is gone.