Error 3343: Unrecognized database format, if 2 front ends open at same time (1 Viewer)

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Hi,

Getting the above error after a few changes of data, but not immediately.

Not changed or done anything different (perhaps a Windows update a few days ago but sure fine since then)

Windows 10 on all PC's. All been fine & 100% for years (never had corruption)
Access 2007. Full ms access version installed on two main PC's. started with test.accde/runtime switch

3 PC's
Main PC #1 Front End is fine & works no matter what
The back end is on this Main PC #1 and listed as local drive d:/ etc etc.

PC #2 is on wired gigabit network & this is solid.
FE points to BE with \\office\ etc etc.

Now if PC #1 program is running and I then use PC #2 I get the error above
Can sometimes start with Error 3040 Disk I/O but straight after 3343.
And can be a few minutes of using before it fails??

If PC #1 FE is NOT running, PC #2 is fine, still accessing the network BE 100% fine.

I have re-linked the BE, compacted & repaired, fresh import of the BE data into a new DB.

Oddly PC #3 is at another location and seems to work fine with PC #1 FE open. This has just a runtime version. The reference all seem fine & the same VB6. I don't think this PC has had the Windows 10 update yet.


As a note & this was a few week back and been fine for a couple of weeks PC #3 was changed briefly and being runtime I struggled to get reference working. It ended up having VB7 (other VB6). This worked fine. The original PC is back now. This set up and been the same & fine for prob 3 years.

I can't think what to try?

All database I have run diagnostics, all OK.

Done the basic windows scan disk.

Is there a way to repair/install without the original CD, bit reluctant to put the old CD in as sure once thought it put some old stuff on. All legit & registered but not done this for years.

Any advise please.
Thanks I/A
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Forgot to say,
When I open the back end directly it Auto recovers no problem, all data fine.

It does feel like a network interuption but as it works fine if #1 FE is not running and looping speed test show 300mbps and no fail.


Bit more info.
Update for Windows 10 Version 1607 for x64-based Systems (KB3150513)
was installed just before the first time I had the failure. I had not noticed the date, the main update was some days before but this single day just before.
Tried a system restore but typically failed for some reason. Will try an specific uninstall of this update if I can.
Uninstalled this update but no help although did not restart PC, as Win 10 don't have easy option not to update I thought not. On reflection as a windows update doubt it, 2 days prieor there were perhaps 20 updates including lots of Office & specific access, nto sure how much I used the DB after this.
So busy at moment don't have time to test & uninstall all one by one!!!
 
Last edited:

The_Doc_Man

Immoderate Moderator, Former MVP, Retired SysAdmin
Staff member
Local time
Today, 03:40
Joined
Feb 28, 2001
Messages
17,309
The "unrecognized format" error most often comes from failure to properly close an FE that links to a BE, leading to some buffer being in an unresolved state. NOTE this is FAR from the only cause; it is just the one I've seen most often.

In all of your discussion, one point was not clear and I need to ask. Does your #1 PC that has the BE locally EVER give you the "unrecognized format" error? Or is it always and only PC #2 that does the URS mapping that ever complains?

A repair installation is relatively safe provided you then trigger a manual Windows Update after the repair but before you try to use it again. Doing it that way would simply force the updater service to reinstall patches for anything that got replaced by the repair op.

Patch KB3150513 does not appear to relate to Office and therefore at first blush does not appear to be your culprit. Further, since you are already on Win10 on all of your PCs, the patch is probably a moot point. The articles I found suggest that this is one of those patches from Microsoft where they install code to pester you to upgrade to Win10 if you WEREN'T on it - but you are already there so for you it might be benign.
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Hi,

Thanks for your reply.

Never happens on local #1 PC even if PC #2 is running the FE (or not so far but think it would have by now)

Not had much chance to test but so far has not done it on #3 PC which is runtime only (No full install or recent updates)

Only error's on PC #2 & does actually error after a few simple data changes but ONLY when the FE is running on PC #1.
There are few local tables in each FE (All FE copies are basically the same but have a few tables I wanted with independent header info. It actually error's also when changing data on one of these local tables AS WELL as the BE, in fact seems little more prone to it.

It does feel a little as if FE issue but as I say is fine if PC #1 not running.

After an error the .laccdb is still there & does not disappear even though I can quit/close the DB but maybe to be expected as it still has technically crashed but seems to close gracefully.

In all the years I have been amazed that I have not had any problems, this is a first.

I did compact & repair all the FE's also.

I can't help thinking the updates are the most likely culprit.

or somehow they are not playing together nicely when both FE's are open but one way?

Reluctant to recompile in case introduces something new btu will soon as getting desperate.

I can't see a simple way to repair install, there is the option to diagnose and test but this does not seem to do much & says OK. I guess I will have to dig out the old CD. I might try first uninstalling as many of the updates as possible that mention access or office.

Trying to do thing incrementally as I know could make things worse.


If I run at the same time, two of the same (different names) FE's on PC #2 both pointing to the BE on the network PC there is NO error.
 
Last edited:

The_Doc_Man

Immoderate Moderator, Former MVP, Retired SysAdmin
Staff member
Local time
Today, 03:40
Joined
Feb 28, 2001
Messages
17,309
bignose2, I understand you look to a patch as doing this, but if it is a Windows patch I would be reluctant to point to it immediately.

The fact that it is localized to PC #2 is significant. Using the URS mapping (\\server\path) is fine, I've used that for years with no issue, so I'd offhand say that's not the culprit.

The fact that after this error the .laccdb file is still present is not an issue because if you later open and then close the database from the #1 system, then open and close it from #2, Access should clean it up automatically.

There is something lurking in the back of my mind about what happens when you have TWO errors but the one reported first leads to a shutdown. I think that in that case, the other error gets "eaten" and you don't see it.

This is just a really wild-arsed guess, but check how much swap space (virtual space) is allocated on #1 & #2. Normally the default amount is enough, but if anyone did any "tuning" and diddled the swap space, you could have a problem.

One of the issues that must be managed when file sharing is that Windows needs extra virtual memory when you have to deal with distributed file locking. Running the same FE under two sessions on the same PC (#2) does not engage the distributed file locking system at the same level as running the FE from two different PCs. The network doesn't get involved at the same level.

Another thing you could do is go ahead and run your app from both PCs, but when the error occurs, you (or whoever gets the error) on #2 must IMMEDIATELY note the exact time of the error. Then open Control Panel >> Administrative Tools >> Event Viewer and look at the application logs and maybe the system logs (one at a time of course). Look for the time of the error and see if the system recorded any events at that time. I doubt that the "unrecognized format" error would be logged - but something else might be.
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Hi,
it's only me who does anything techy with the computer, just a few other users I know would not touch anything else.

Page file pc1 4096 Windows control, pc 2 8096mb, both pc well spec's but is only 4gb ram on pc 1.

Its just I think too much of a conisidence that a few days after lots of updates, office included, that this has happened. Everything else is fine.
As I say I have had this exact setup for years and we'll over a year with Windows 10.

Really struggling here, out today so no chance to investigate, can't easily install office on another pc to test but think I might have to find a way.
 

The_Doc_Man

Immoderate Moderator, Former MVP, Retired SysAdmin
Staff member
Local time
Today, 03:40
Joined
Feb 28, 2001
Messages
17,309
At first blush I would review the Office updates before I would point to Windows updates, since your description suggests you don't have a lot of time for detective work.

Having MORE space on PC #2 eliminates virtual memory problems because I can almost guarantee you that nobody diddled with a profile's max virtual memory number. That is only used very much on CITRIX-like server environments or database server environments where a single thread is unlikely to do much so the default virtual limit is set down to allow more stuff in memory at once. In your environment, I don't see that happening.

However, the fact that ONLY PC #2 ever throws the error still is significant. I still think looking at event logs for the date/time of the error is a valid source of forensic data. You still need to know what happened even if it wasn't a virtual memory error. When you look at the logs, if your timestamp for the event is X, you can probably limit your search for that time between X - 1 minute and X + 1 minute. Outside of that, you are not likely to see much.

Don't forget to check network logs, though. Don't limit yourself to system or app logs. If you DID have a momentary drop, you might see an "unrecognized format" error because of that. You are using SMB protocol, which is a TCP protocol (vs. UDP) and therefore has TCP/IP connection verification safeguards that would be disrupted by a momentary network drop. The event log might show Device Up / Device Down errors for your network device, whatever it is, and that would tell you the story.
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Hi,

There is nothing in the logs to indicate any issues or errors.

In frustration I thought I would upgrade all to Office365. Really did not want to but after most of the day trying anything I can think of run out of ideas

However much to my horror (little strong) it has not helped!!

So far only installed on PC #2, the one with the problem and it still go the error. Sure network OK, really fast & I say works when PC #1 is not running the DB.

So really at a loss now and even more desperate, unless it is with PC #1 the BE PC but I don't see that.

My thoughts are PC#2 had Windows or hardware issues. Strange actually as on the 15th there were 20 or so updates, then again yesterday there was another large Win 10 update (Being so soon , perhaps a hasty fix) but as it did not help??

I have run a sfc /scannow nothing to report.

Really really don't want a Full re-install, a huge undertaking with all the stuff I have set up.

Not a hint of any other troubles.

The only thing I have posted on Windows 10 forum which now I think is more significant, a few folders in the PF (X86) MS shared folder are hidden & I can't unhide, I only saw them because they are visible in the references browse, really off, but I figured would error or say missing if not actually using them.
The only other thing I am just about to try is MalwareByes, I have installed, I am sure it was listed on a restore point about a big update. DOn;t know what it is with Win 10 as only 2 points listed at start of my problesm, would not restore to eitehr & now none are listed. Over the years win 7 & 8 reliably saved loads of point and could revert without ever a problem.
 

The_Doc_Man

Immoderate Moderator, Former MVP, Retired SysAdmin
Staff member
Local time
Today, 03:40
Joined
Feb 28, 2001
Messages
17,309
I can't unhide,
If you can't unhide the folder, then you are not running as administrator OR (possibly) you have been infected. But if you did the installation, you had to be running as Admin or else that would have failed. This is a contradictory set of symptoms. Do you have good anti-viral software on PC #1 and #2? If so, does a full scan reveal anything?

As a side note, if going to Ofc365 didn't help and you have a way to do this, I would step back. OK, economically might be an issue, but a year from now, you'll get a bill from Microsoft to renew your Ofc365 license. With your prior version, you wouldn't get a yearly bill.

I'm still not convinced that this kind of behavior would be caused by Windows updates. If you can make it happen, is it possible to swap the roles of PC #1 and PC #2, so that #2 becomes the BE host and then see if the problem moves to #1 when #2 is running?

Are there any functions performed on #1 that would not be performed on #2? Like for example some kind of cleanup macro? There has to be SOME difference between the two? Check the FE settings under File >> Options >> Current Database for both and see if anything is different? Like maybe one has "Auto Compact & Repair" and the other does not?
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
Hi,

I really appreciate your efforts here.

I think & hope it is MalwareBytes!!

In case I have just been lucky for 10 minutes to cover your points.
I did think virus's. I run a full defender scan & it was OK but it was this that made me think Malwarebytes as I like to run this also & it is in the background, never had a problem before but it made me remember an update was one of the restore points (which soon disappeared). Whilst running the scan (which was OK) I was thinking perhaps this could be the problem.
I really thought not as everything else works so well, I figured had to be access & the sharing & BE etc. I did think AV briefly as is often the case, but just so specific to the DB.

I think still off the hidden file thing but oddly just think not related in this case but would like to undertstand. I can see lots of other files & folders so it is a select few folders I can't see which is odd.


I am still hesitant to light the cigar as after installing the new Office365 it was quite a few operations longer before it failed, I had thought fixed, but a few times before it was slower to fail.

This time though really been quite a while but need a break now.

thanks again, will keep posted in case others have similar.
 

The_Doc_Man

Immoderate Moderator, Former MVP, Retired SysAdmin
Staff member
Local time
Today, 03:40
Joined
Feb 28, 2001
Messages
17,309
Here is the issue that I see.

You have two machines on the same versions of Access and Windows. One machine hosts the BE, the other links to it. When the hosting machine and the linking machine run the app at the same time, the linking machine barfs. That tells me that SOMETHING is different between the two machines. And re-reading the lead-in post, I realized I actually DID know of something different between the two.

I have one more crazy idea for you. Don't run the #1 machine's BE locally through the D: drive. Link the BE to the #1 FE using URS mapping just like the #2 machine has to do. Doesn't matter that the URS mapping will be resolved in-machine. All that means is that the network software will perform network operations over the Loopback circuit, which is a thousand times faster than the network because it never leaves the machine. BUT it will go through network lock management and BOTH machines will be forced to be nice to each other with respect to locks.

The difference between the two machines is that URS mapping ( \\server\path...) and local mapping (D:\path...) use DIFFERENT SECURITY RULES! Local mapping and network mapping resolve differently because domain rules (sometimes) don't apply to local paths. The access arbitration software that is part of the lock manager sees a local path and does not have to check the domain (as much). Or maybe the domain never even gets involved. What is the difference in the rules? Damned if I know, it is specific to each site's configuration rules and automated startup policy, but there frequently IS a difference. Put the two machines on the same footing and see if the problem repeats itself.
 

bignose2

Registered User.
Local time
Today, 09:40
Joined
May 2, 2010
Messages
176
SOLVED.
Hi,

Thanks again for all you help,

Just to confirm that the problem was caused by Malwarebytes Anti-Malware.
I uninstalled and all is fine, I could not be sure until I had a good day to test.

I know perhaps I should have considered AV from the start but it was such a weird & specific problem and coincided with all those Office & Windows updates. I think if it had failed all the time & not just when the other PC was running the DB I perhaps would have not been so single minded towards MS.

Had for years and never had even the slightest problem. Seems they have loads now!

It was only later I noticed there had been quite a large upgrade to Malwarebytes just before all my problems

Really annoyingly is Windows 10 restore seems a little unreliable as if it had restored (it failed) before this point I could have easily narrowed it down and not wasted 2 days

Fortunately New Office365 installed & works well (touch wood) and not bad to have an upgrade from 2007.

Thanks again.
 

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top Bottom