FM10 Seemingly Random Crashes

[hermanpeckel]

Folks,

I'm having an absolute shocker here with users reporting Filemaker crashes throughout the day. We are running Filemaker Server 10 on a Win2K3 SP2 platform and we have a mix of clients on OSX and XP. Three of the windows clients are having issues where they do seemingly unrelated things and Filemaker will crash. The error messages received are -

"An error occured during this operation" (very descriptive) and

"An unsupported operation was attempted" (again...)

Also we get the generic "Filemaker Pro.exe has encounted a problem and needs to close", windows thing.

It is happening on three PC's, the other 40 or so are ok. I have tried the following things to fix them -

Removed and reinstalled Filemaker (including deleting the original program file folder)

Updated the installed filemaker with 10.0v3 (the rest of the machines are on 10.0v1)

I have also patched the machines to a (fairly) similar level of a known working machine. I even pulled the HD out of one of the buggy machines and put it into a similar machine (cause the original machine died), but this still bugs out.

I'm kind of at a loss now as to what else I can try. If you have any thoughts or suggestions I'd love to know!

Thanks in advance!

HP

[Vaughan]

First make sure that the client machines are updated to the latest version. There should be no client running 10.0v1 it has some nasty bugs. Also check that FMS is up to date and patch it as necessary.

Also check whether third-party plug-ins, if used, are up to date on each machine. Be aware that there are several locations that plug-ins can be installed so check all.

Consider that the database could have corruption. Get a recent backup and take it off the FMS and run the recover command over it, then check the recover log for issues.

The "unsupported operation" errors are interesting: could it be possible that the file has FMP 11 features in it?

FMS and FMP should be very, very stable. It should not be crashing and should NEVER require force-quitting.

Try increasing FMP's cache (in the application preferences) on each client machine.

[hermanpeckel]

Ahhh!!! Thanks for the response Vaughan!!!

You bring up some very interesting points. I will check all of the options you have mentioned, the cache size being the first as it will be pretty easy!

So, if the database has corruption, would you not expect other users with similar roles would experience the crashing as well? As far as I know everyone else is running sweet.

I will check with our FM dev guy and ask him about possible FMP 11 features. This is a distinct possibility as we are currently preparing to upgrade to 11 or 12.

Thanks again. I now have a glimmer of hope to cling to again :laugh2:

[Vaughan]

You would't upgrade to FMP 11. You'd go to 12.

If you're doing that then I'd recommend setting up a development server and doing test conversions to assess the impact and the issues that will arise.

[hermanpeckel]

So far Vaughan, increasing the cache size has worked, but it's early days I guess. I upped it from 8mb to 16mb.

Thanks for the advice and I'll post an update if it remains stable, otherwise I'll look into the other options.

Cheers!

HP

[Wim Decorte]

I've seen behavior like this on underpowered workstations: either not enough free disk space on the C drive or not enough RAM for everything that is going on (not just FM, other apps as well). Most notably I have seen this with terminal emulation programs (for AS400) with multiple windows running together with FM.

[hermanpeckel]

So, I now have the need to revisit this topic (unfortunately). After much mucking about, we ended up resolving this issue by buying three new PC's. The way I figured it was the machines in question were about 5 years old, and well and truly due for upgrade anyhow. Now we have three new machines all running sweet. BUT there are still a few old ones about and these old machines are now having this problem.

I have tried the above suggestions - FM client update, OS update, FM reinstall, checking versions of plugins, basically everything above. Also there's a registry fix I tried - http://www.google.com.au/url?sa=t&rct=j&q=&esrc=s&frm=1&source=web&cd=2&ved=0CCYQFjAB&url=http%3A%2F%2Fhelp.filemaker.com%2Fapp%2Fanswers%2Flist%2Fsearch%2F1%2Fkw%2Funsupported%2520operation&ei=QPpGUI3fDc6fiAes7YGYAg&usg=AFQjCNFjPpYN7QL5bHUZ_oiIwxXKIYIXNw

No real change.

Wim, you bring up a very good point - all the machines that have been having issues, although C drive and RAM have been ok (100G spare, 2GB RAM), they were certainly old bangers. This is the case with this current issue.

Does anyone have any other suggestions?

Regards

HP

[Ted S]

Back in the FM7 Win XP days I had a couple PCs that would crash whenever the user moused over a 1px by 1px line. It didn't happen on every PC, but on a couple I could make it happen at will.

I solved the problem in two ways:

1. I got rid of all 1px by 1px lines.
   
2. I replaced the video cards/drivers in the old machines.
   

[hermanpeckel]

Yeah, I'm beginning to think the only way I'll solve this chestnut is to replace all of our PC that were made in the 70's. We've never had the issue on newer hardware. Ugghhh...

[hermanpeckel]

Hmmmm... just revisiting this AGAIN. The issue has popped up on another machine here. I'm wondering if the issue may be related to this -

http://help.filemaker.com/app/answers/detail/a_id/7644

EXPLANATION:

The registry for toolbars increments 2 entries per launch. When the application has been launched and quit ~1000 times, FileMaker Pro will no longer launch.

I have run the batch file on the users machine and will post if that seems to fix it. If anyone has any other suggestions I'm very open to them. This is like the issue that just won't die!

Regards

HP

[MarkThomas]

Forgive me for necro-posting, but it took me quite a while to find this so it might be worth a bump.  We have a system running on FM10/Win7, and were having major problems with these unsupported operation errors which at some times were causing a corruption requiring recovery.  At first we thought it was because one of the tables had grown to excessive size (several million records)..  

 

But after finding this thread and a couple of other references to that registry problem I, like the poster above, used the batch file referred to here to delete the bad registry entries.

 

The errors vanished.  We still need to deal with the excessive table, but life is now much easier...