6 users 3 no problem 3 many problems

Check your references...it might seem the same...the correct ones...the ones you use all the time..but it might not...check even the sub version of the .dlls you are using...i had a case a tough case years ago where a .dll had the same version...looked pretty much the same but a subversion difference would cause real havoc.

Hi John

Are the Dlls part of an installation (In my case Sagekey software install everything as far as I know) or are they already in each pc and called by an application?

maybe you should check...also mixed 32/64 bit...seems like a good candidate for issues.
Probably ...since windows 7 works fine ...you are utilizing some 3rd party ActiveX that doesn't play well with Windows 10 and sagekey ...maybe...just maybe..."hides" the problematic misregistration

What's on the server?   and have you disabled SMB leasing?

Jim.

Sorry, I missed this:

We got over the issue of "The database is in an inconsistent state" and have applied the patch. This has not been a problem since and the installation now reports very very rarely does an issue require a reboot.

 So you probably had a legit network error.   Is the server virtualized?

Jim.

Is the network wired or wireless?  Access is unstable on wireless networks since the network frequently "blips" and Access cannot recover.  It would be more stable with a SQL Server BE but I still wouldn't trust it.

Hi Pat  There are no wireless connections but they do access the data (cable) from another access program (16) but it is only linked the same as the app.
Jim, Virtualized? I can ask the question. What does that mean as far as access is concerned.

I will add too that with a good network, JET/ACE is quite stable.   You can easily run up to 30 or 40 users without issue.

Jim.

Thanks Jim that's very helpful. I can send that to the IT engineer and put a little pressure on them for a change.

Is there anything here that would differentiate between windows 7 and windows 10. This is still a central mystery.

We have disabled SMB leasing and we think that this has stopped us having to compact and repair the data file two or three times per week. No repairs have been necessary since.

Between the 6 users, my app crashes 12-15 times per day.

We have disabled SMB leasing and we think that this has stopped us having to compact and repair the data file two or three times per week. No repairs have been necessary since.

 So what  do you mean by "my app crashes"?  I was getting the impressions that C&R's are still required on-going and this was networking.

 Also, what version(s) of Access?

My app is an Access mde packaged and installed using SageKey software using Access 2002 runtime. We get disc or network error 9 times out of 10 but the other 10% Access closes down by itself.

6 users, 3 on windows 7 no issues, 3 users on Windows 10  all 3 have office 365 of those 2 are 64 bit and one is 32 bit

The back end data is an Access 2002 .mdb file.  The client accesses this data with another database (Access 16) to gather info for other application (Excel)

Windows 10 doesn't play well with very old versions of Access..i know for a fact that W10 and A2k3 works for running the application but not for design

Are you using mapped drives or no?

Thanks John.

Jim We have a re-linking procedure that sees that the data is not where it is expected to be, as it would on a single user PC in the installation drive. It then opens a windows file explorer asking the user to direct our application to the data on the server. It then re-links and opens the main form. This has never failed even in the current user environment. Is  this a work around for mapped drives?

<< Is  this a work around for mapped drives?>>

  It depends.

  When you use the window to browse, do the users pick a drive letter?   Or do they go to network, find the server, then the share?

  The first is a mapped drive, the latter UNC (Universal Naming Convention  - \\Server\Share)

  The reason I asked about drive mapping is that there is a setting in windows to disconnect inactive connections.  So if your users are sitting for a while, then go to do something, the connection needs to be restored and you can see network errors.  

 The other issue for mapped drives is a GPO (Group Policy Object) that is set to "replace" instead of "update".  Replace forces a disconnect of the mapped drive, where update does not.

Jim.

"When you use the window to browse, do the users pick a drive letter?   Or do they go to network, find the server, then the share?"

The former. The windows explorer handles it by clicking drive the file the share

Would this replace instead of update be applied to each connection (PC) or to the data file?

<<The former.>>

 That's a mapped drive then.

<<Would this replace instead of update be applied to each connection (PC) or to the data file?>>

  It's applied to the drive mapping if being placed there (on a station) by a GPO (Group Policy Object).   For mapped drives, you can set them up on a users station via a login script (that's old school), or by the admin using a GPO.   In the GPO, you can create, replace, or update a drive mapping.

 Replace is often selected because it works if the drive mapping is or is not already there.   Create only works if it's not already there, and update expects it to be there already.

 The problem with replace though is that active connections are closed, the mapping deleted, then created anew.   GPO's by default are applied every 90 minutes.

 You can avoid all this by switching to linking the tables with a UNC path, but that might not be possible for various reasons.   If you don't, then you need to check if mappings are being setup via GPO, and also change the timeout on the server and stations to the max value so they don't timeout.

I can go into details on that if needed.   Pass this along to the admin folks and they will understand.

Jim.

I have just updated the back end data to an accdb on my own pc. do you think that would make any difference at all to this issue?

Strangest thing I linked to the BE Access 16 accdb with my 2016 accde all perfect except for one report whos query had the forms item number on it set to invisible. The data in the reports query had no records. I set the "item number" on the main form to visible checked that it had a number when in form view changed it back to invisible and the query now shows data. I wonder sometimes if I am going insane!

Would not be a problem but if I change the back end on the clients server to accdb and then all pcs have the same problem I am going to look pretty stupid

<<I have just updated the back end data to an accdb on my own pc. do you think that would make any difference at all to this issue? >>

 No.  In terms of networking and how the database engine functions, there is little change between a .MDB (JET) and .ACCDB (ACE).

 The only real changes have been a few new data types in ACE.   Outside of that, they function and operate identically.

Jim.

My thanks to you all again.

This unusual recognize experts form, I thought it was some sort of point system but I see that if someone has contributed and you want to recognize them by clicking say 1/4 bar that it may be indicating that they have not been helping as 1/4 is not far from Not helpful at all. Presumably indicating that the contributors have helped would then necessitate a higher rating. Does this effect point allocation or has all the points thing stopped.

The points are done behind the scenes now.   All you need to do is select which comments are helpful or not, and which one you feel was the solution.

You can see who was awarded points in the upper left by the question.  And you can re-assign who got what by just clicking on the buttons in the various comments.   There is no question close process anymore.

Jim.