Access Database is getting corrupt again and again

Starting 1 year ago I have had the same problem on several sites with the same database schema. In all cases the database sizes were around 1.2G. Repair was simple - opening with the full version of access auto fixed the "corruption". But "corruption" kept happening with users getting the "Unrecognized format" error reported. After trying everything else I suspected that for this particular database, the size was triggering the error (I know that the max size is 2G and on the same computers there are several other databases being accessed simultaneously that are much larger than 1.2G that do not corrupt. Some of them are 1.8G with no problem. The only difference is they have less complex schema than the one that corrupts.) 

As soon as I cull enough records to get the databases below 1.2 gig, say 800MBytes, the problem goes away. Not ideal to have to cull records but the customer is at their wit's end with the disruption. Over time when the size creeps back up to 1.2G the same corruption happens again and again. 

So in some cases where the schema of a database is above a certain complexity, the database size limit seems to have reverted to be around 1.2GBytes. All the errors are the same as those encountered when the size exceeds 2G.  I know the max size is supposed to be 2G and I have lots of sites with databases between 1.2 and 2G that have no problems. But this one specific database, with a complex schema, ALWAYS corrupts when the size gets to 1.2G and is always fixed by reducing size (unfortunately culling useful records). 

After 10 years of no such problems this has started to happen in the past 12 months. It is not limited to Win10 - it happens on Win7 as well. Something has changed in the OS  - whether the database is in .mdb or .accdb the problem is the same.

Has anyone else experienced this 1.2G size limit?

None of my databases have reached that size yet so I can't comment on the size issue.  I do know though that the five sites that have implemented the temporary fix below are no longer having problems.  (Fingers crossed.)  Waiting for a PERMANENT solution, Microsoft...

https://support.office.com/en-us/article/access-reports-that-databases-are-in-an-inconsistent-state-%EF%BB%BF-7ec975da-f7a9-4414-a306-d3a7c422dc1d

We had the same problem - multiple database corruptions per day caused by PCs with Win10 1803 - but applying the registry fix for client computers running Win10 1803 went all wrong: although we did not experience database corruption anymore, performance of Access application degraded drastically to almost unusable state (mostly on brand new machines with i5 8500, 8GB RAM, 240GB SSD). 

Our setup:

Server hosting the BE with Windows 2016, client computers with Windows 7 or Windows 10, all of them on a gigabit network.

The next thing we tried was reverting registry changes of Windows 10 1803 PCs to default values, and applying DisableLeasing on the server instead. Performance instantly increased to almost the same level as was before fixing anything, haven't had (fingers crossed) any database corruption for two days now. To my knowledge, we did not experience any side-effects of DisableLeasing so far.

But this nasty bug has to be fixed. It is replicable, well described with lots of evidence, yet it takes too long to fix.

Thank you!  I'll bear this in mind and check to see if the users are reporting any degradation of performance.

Thanks for info. The problem on the Win10 version is probably related to the problem I have described but the fact that it is always fixed by reducing the size below 1.2G and this occurs on both Win10 and Win7 hosting the data means applying this would not fix the problem. I was hoping that by posting the details that it might alert/help the Microsoft Engineers working on this problem in finding a solution.