We help IT Professionals succeed at work.

Access 2013 - Access error “Your network access was interrupted”

Hi All,

I am having a nightmare of an issue with some of my access databases. We have around 20 different access databases. The larges database has around 30 users.

Everything was running without any issues until we swapped 150 users to Windows 10. We had tested our databases within windows 10 but had no issues.

We are running W10 build 1903, the databases are sitting on a network share using UNC paths not mapped drives. Our file servers are running on server 2012.

around every 20 min users will receive an error message “Your network access was interrupted”. This is happening to our whole userbase so around 150 users.

I have ran a ping command to the file server to see if the network drops when we receive this error. The pings are stable at 1ms and we don't see a drop in the ping when this happens.

If anyone could shed any light on this it would be massively appreciated, thank you in advance.
Comment
Watch Question

Distinguished Expert 2019

Commented:
You should see if the share access (smb=port 445, not ping!) is stable. Use something like portping.
Phillip AshcroftTechnology Manager

Author

Commented:
Thank you, could you provide a link to download portping as I am having difficulty finding a version I can run. I have tried

https://sourceforge.net/projects/portping/

Thanks
Distinguished Expert 2019

Commented:
I was suggesting portping by Thomas Kolb, but Thomas' site is down.
I could upload it here, but it's an executable, so I will rather offer a link: https://send.firefox.com/download/39975258e20a0f1a/#vI9AZMgdDjxQ_6BrXr0Snw
password: see my personal message
Daniel PineaultPresident / Owner CARDA Consultants Inc.
Distinguished Expert 2018

Commented:
Do you establish a persistent connection at the startup of your database?
They are all connect by a wired LAN?

You are aware of this ongoing issue caused by Windows 10?  
https://www.devhut.net/2018/06/13/access-bug-database-is-in-an-unrecognized-format/
Dale FyeOwner, Dev-Soln LLC
Most Valuable Expert 2014
Top Expert 2010

Commented:
1.  Daniel's comment about "wired LAN" is critical.  If anyone is accessing the BE over WiFi, you are likely to get corruption of the BE database, with Access, I encourage you that all users should be using a hard-wired connection to your network.

2.  Does everyone have their own copy of the database?  If not, this is the next critical issue that you should confirm.  None of the users should be sharing the application file across the network.  Everyone should have their own copy running on their PC.

3.  Is this a new situation (within the last week)?  What version of Office are your user running?  There is a new (on or after Jan 8, 2020) bug introduced into Office which will cause Access not to recognize Identity Columns (autonumber) in linked SQL Server tables, which will cause problems whenever you attempt to write to tables with an identity column  (more from Microsoft on this)
Phillip AshcroftTechnology Manager

Author

Commented:
Hi all, thank you for your input.

I will try portping and report back.

All computers are using a wired connection, no wifi connections are used due to database corruption as you had mentioned.

All users have their own copy of the database on their machine.

We are running office 2013 we have not made any changes to office. We have upgraded our machines to windows 10 build 1903 around 150 machines in total. I have tested access 2016 and oddly this seems more stable on windows 10 than 2013 but this could just be coincidence. Access 2016 still has the same issues but not as frequent.

I was not aware of the new issue.  I was aware of the December bug but not the Jan bug. I will have a look at this.

Thank you.
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:
It could be caused by malfunctioning or misconfigured network equipment - a switch.
To rule this out, discuss with your networking guys how to temporarily establish an alternative route for, say, a dozen workstations to the server.

In line with this, consider if there were any network related difference between your successful test with 10 workstations and the current setup.
Phillip AshcroftTechnology Manager

Author

Commented:
Hi Everyone. I am back in the office now and will start to try a few of your suggestions this morning and report back later on. Thank you all so far.
Phillip AshcroftTechnology Manager

Author

Commented:
I have run a few tests to far with portping on clients machines. I can see that on port 445 in the main is a 1 ms response time. there are a few spikes up to 16 ms but only a small handful.

I have checked out notes from testing our W10 computers and nothing has changed in terms of the network config.

Does anyone else know of any other Database bugs that are currently ongoing forAccess 2013 or 2016?

Thanks.
Distinguished Expert 2019

Commented:
Look, what portping says while it is working, is of no relevance. Let it run all the time and see what it says while the problem happens.
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:
There should be none that can cause this kind of error.
Distinguished Expert 2019

Commented:
Apart from this, diagnosis should be easy as this happens every 20 minutes. If it is indeed, as suspected, buried in how windows 10 works (I don't think so, but we'll see), then setting up win7 now for a test should clear that up pretty soon, shouldn't it?

So please set up win7 to compare with.
Phillip AshcroftTechnology Manager

Author

Commented:
Not sure if this is relevant or not but I've just noticed that whilst saving a word document to the same server from within word there was a little lag on connecting to the server. Port ping did Jump yo 15 ms but only the once. The database running on my machine did not crash out.  Seemed a little odd so thought id mention it.

I still have 20 windows 7 machines on the network and they are all working fine without any issues. I will run porting now to see if I get any spikes as I'm currently seeing on our W10 machines.

Thanks
Distinguished Expert 2019

Commented:
Little spikes may occur (once every 100 packets, maybe) and shouldn't be a problem. I have that sometimes, too, and the lag is about the same as yours (below 20 ms).
Phillip AshcroftTechnology Manager

Author

Commented:
Everything on Windows 7 is really stable, no crashes with the databases, portping is reporting a consistent 1ms response time.

On the W10 machines, we are seeing stable 1ms with the occasional 15ms spike.

W7.PNG - Windows 7 Machine

Capture2---DG.PNG Windows 10 Machine

When the database crashed on the W10 machines, there is a small spike to 15ms but this is not consistent. Sometimes it will crash and not spike to 15ms.
Captpingport.PNG
Capture2---DG.PNG
Captpingport.PNG
Capture2---DG.PNG
Distinguished Expert 2019

Commented:
Just so I understand: these two screenshots, which one is win10 and which is 7, why did you show us 2, since clearly, both have spikes.
Phillip AshcroftTechnology Manager

Author

Commented:
Apologies uploaded the incorrect images, I have edited my previous comment. please see the above.

Thanks,
Distinguished Expert 2019

Commented:
Hmmmm, I did a test and fired up a win7 VM (we got rid of all win7 years ago). portping shows zero spikes, none at all in 300 packets (=5 min), while win10 (several I used this on) show occasional spikes, all at 15 ms. I wonder why...

Nevertheless, we don't have instabilities here with office files of any kind.
Phillip AshcroftTechnology Manager

Author

Commented:
I thought the same, it seems really odd that W7 is so solid and W10 shows spikes, I'm not sure either why this is. Good to know though that this is normal behavior for W10 to show the odd spike.

I will have to try and have another look to see what could cause our issue.

Thanks
Distinguished Expert 2019

Commented:
Normal behavior? That is not clear. It could be that all my test machines (and yours) have the same NIC driver which might have a problem, who knows.
Phillip AshcroftTechnology Manager

Author

Commented:
At this point now, I am going to upgrade an old W7 machine to W10 and then run the same tests to see if we still have the same issue. If we don't then we can possibly assume its something to do with the new computers.
Phillip AshcroftTechnology Manager

Author

Commented:
After upgrading on old HP Windows 7 machine to Windows 10 I am getting a solid network performance. After using the portping util I am now getting a constant 1ms response time with no spikes at all. I have run one of our databases overnight and it's not crashed out at all.  

All our new computers were built and imaged by a third party, I am now going to restore one of the new machines to factory settings and rebuild the image and then test again. I am hoping this will solve the issue.
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:
Great news.

At this stage I would suspect the NIC or its driver in the new machines.
Notice if the restore will apply a different driver.

Such issues are rare these days, but I recall an issue with some specific Dell machines and a Broadcom driver.
Phillip AshcroftTechnology Manager

Author

Commented:
Looks like the issue was caused by an out of date driver. We have updated the driver on 5 machines and are awaiting results over the next 24 hours. if we have no dropouts o these machines then we will roll out to all machines.
Phillip AshcroftTechnology Manager

Author

Commented:
Hi all after further testing and rebuilding of machines it turned out to be an old intel driver that caused the issue.

thank you for everyone’s input.
Technology Manager
Commented:
Hi all after further testing and rebuilding of machines it turned out to be an old intel driver that caused the issue. i

thank you for everyone’s input.
Most Valuable Expert 2015
Distinguished Expert 2018

Commented:

At this stage I would suspect the NIC or its driver in the new machines.