Hi,
Please find listed below the following articles & suggestions which I have compiled over the last few years & use to overcome stability problems that I encounter, which may be the cause of what you are experiencing.

The Dataaccess article covers OPLOCKING & READCACHING, whereas I have added the disabling of the write caching if required.

The following should be implemented progressively in the following order until you get stability.

1. Howto disable autodisconnect http://support.microsoft.com/default.aspx?scid=kb;en-us;297684&Product=win2000

This applies to all Server versions, and especially to sites where the Workstation edition has been installed on the Server in lieu of the Server Version.

The number of shares on the Workstation version is limited to 10, this and Autodisconnect can cause problems in the following scenarios.

a. The default timeout is 15mins (f hex), if the connection to the server is cancelled, and due to load conditions on the Server and it is unable to respond quickly enough to the client request we receive a timeout from the Database. Once in this error condition occurs any subsequent reconnect is suspect and Database corruption will occur.

Please Note: To disable autodisconnect is not 0, this will cause disconnection after a few seconds. The correct setting is
ffffffff Hex i.e. 4294967295 minutes.

b. If the current number of shares is 10, to a Server running Workstation edition, and an autodisconnect occurs, then
another process takes the remaining share, we are unable to regain a connection to the database. i.e. database corruption

Please note i. From testing I have found that the current number of calculated shares can be less than actual.
ii. Although the article says that the maximum command line value in minutes is 65535, the registery accepts ffffffff which is considerably larger.

2. Dataccess recommendations http://www.dataaccess.com/whitepapers/opportunlockingreadcaching.html

3. Disable write caching http://support.microsoft.com/default.aspx?scid=kb;en-us;324805

4. Disabling write caching may not stay disabled http://support.microsoft.com/default.aspx?scid=kb;en-us;290757

5. Some SMB Hotfixes causes problems with SP3 http://support.microsoft.com/default.aspx?scid=kb;en-us;326798&Product=win2000

6. Windows Redirector may leak file handles http://support.microsoft.com/default.aspx?scid=kb;en-us;289509&Product=win2000

7. The Windows Redirector May Cancel a Session During a Long Locking Operation http://support.microsoft.com/default.aspx?scid=kb;en-us;320310&Product=win2000

8. During a OEM Preinstall the GUID is generated, but it can be a Duplicate of other Workstation GUID's, is will cause OPLOCKING to fail, so best thing to do is disable OPLOCKING because of this and the other problems that are experienced.
Note: This does not occur during a normal install as GUID generation process checks for Duplicates on the network.

9. RFCB caching should be disabled, as it causes problems when a file close is issued, the OS signals that it has closed the file, but it is only done in the cache, not to disk.

Under load conditions any subsequent file opens will result in spurious results, which eventually lead to database corruption.

This affects many other 3rd part products as well as MS products i.e. excel & access.

The link below describes how to disable.

Disable RFCB caching http://support.microsoft.com/default.aspx?scid=kb%3ben-us%3b126026

>Hi all,
>Is there a function that forces VFP to update its cdx files since flush only updates table header and not cdx header?
>Thanks in advance
>Yau

Hope this helps.

Regards N Mc Donald