Back End Crashing

[Eljefegeneo]

I work remotely from the main office, mostly retired but helping out the Access Databases. For some time now the back end of the DB has been crashing with regularity. We finally narrowed it down to one computer terminal that seemed to be the culprit. For months I have been telling them to check the cable connection to the server. They kept telling me it was solidly connected. Yet the back end still crashed when this particular terminal was using the Access DB. We installed a new server and this did not solve the problem. I changed the mapping of the back end link but no joy either.

Finally they told me the following. "We rerouted the network cable so it was not going through the VoIP phone." Since then (and I hope I do not jinx it by relating this tale), the problems have seemed to vanish.

I pass this bit of information on since it might just help someone else experiencing the dreaded back end crash.

 

[pbaldy]

Thanks for the post, it will likely help somebody down the line. Good job being able to isolate the problem computer!

 

[isladogs]

Thanks for the info. Another variation on the issues that can occur when users have a WiFi connection to the network BE

 

[Minty]

Interesting. We had an issue that some of the older IP phones throttled the passthrough connection to 100Mb, (less whatever the phone was using during a call), which caused us some headaches, and was really noticeable when we had spent time and effort putting in a nice 1Gb Cat 6 network in the whole building.

 

[Deleted member 147267]

Another variation on the issues that can occur when users have a WiFi connection to the network BE

I interpreted the post that the pc was hard wired?

 

[isladogs]

I interpreted the post that the pc was hard wired?

So did I.
My point was that any kind of interrupted connection can cause problems of this kind whether due to a dodgy VoIP phone or loss of Wifi connection.

 

[Deleted member 147267]

Point well taken!

 

[Eljefegeneo]

Spoke too soon. Worked for three days and now getting the message: Disk I/O Error During Read. New I would jinx it. The computer is hard wired into the server. So, either there is too much running in the background and not enough memory, or the LAN port is bad, the cable is bad, or ?? The computer is fairly new, maybe six months old, Windows 10, Office 365. They are going to add memory and see if that fixes it. More to come on this. The server is brand new and the other three computers are OK - no problems.

 

[Deleted member 147267]

Are you 100% sure that it can't be related to the user ( e.g. a profile or NAS permissions issue). If you swap pc/user combo the problem might persist with a new user on the pc or vice versa. I doubt very much that the problem is insufficient or unstable memory.

 

[Eljefegeneo]

Thanks, I will pass that suggestion on to our outside tech.

 

[shadow9449]

Is the power setting on the network card set to turn off when idle? That's a pretty common setting and can be similar to yanking your LAN cable out while Access is connected to the database. You can check from the Device Manager > Network Adapters and then find your card and click the Power Management Tab.

 

[Minty]

What @shadow9449 said. We had a machine (Laptop) set to sleep even when on external power, caused Access to crash at least once daily.
The user only reported it after about 3 months, said he thought everyone was suffering from it.

 

[Eljefegeneo]

The crashing occurs when someone is using the Access DB. But I will pose the question to our Tech gu anyway. He sent me the following two screen shots if they are any help in diagnosing the problem. Much appreciated the help so far.

 

[Isaac]

That oft-stated problem with users who are on Wifi, bothers me--This seems like something that the software & hardware makers of the world are VERY late to fix. I mean c'mon, this is the year 2020........Rarely are people on home computers NOT using Wifi (and those who aren't it's probably because they made a conscious choice for hardwiring because they knew and appreciated the seriousness of the issue). It is primarily "the" way people access the Internet.

I find that the answer, being, Tell all users that they must find a way to stretch an Ethernet cord from wherever their office space in their home is, to wherever their router/modem is, to be woefully insufficient. I did stuff like that when I was a younger and cable internet was just becoming widespread. I'm not denying it's true, but I admit I had trouble accepting it when told this over the last few years. How can something that is the "modern standard" of being online, be "widely known to cause major problems" and no way around it but wired?
This is messed up Microsoft

 

[Eljefegeneo]

Unfortunately they are all hard wired. So goes on the search for the missing clue.

 

[The_Doc_Man]

Has anyone looked at the back-end host's error logs for the time of the crash to see if you got an error out of it? And I would also look at the error logs for the system hosting the active front-end at the time of the crash. If there is a proper entry for the BE file, we might get a clue on it.

 

[Eljefegeneo]

OK, the following was done today:

1. The power settings were checked on the network card again which were off but after digging a couple more power settings that could cause issues were disabled.
Set Auto Disable Gigabit to Disable.
Set Green Ethernet to Disable.
Set Energy Efficient Ethernet to Disable.

Also reinstalled the driver from Dell as I noticed some disconnects in the event logs.
Cleared all the logs so they can be review them the next time it happens to see if there's anything there that can point to the issue.
Also installed the Realtek Ethernet Diagnostic Utility. This will show if there is any errors during the time the user is on Access.

Won't know until next week so stay tuned.

 

[The_Doc_Man]

Sounds like you have a reasonable setup there to trap errors. Just remember that the system error logs are sorted chronologically, so you should not have trouble finding any errors as long as you have a way of knowing when they happened.

 

[Eljefegeneo]

It's been four days since the changes and there have been no problems. Actually I'm posting this to see if my jinx factor holds. But the one user with the terminal has been using the Db since Monday and no problems! May be too early to say, but thanks for all your suggestions. It definitely prodded the outside Tech to look at them.