2x Application server corrupting dbfs

Is there anyone who's got experience of Terminal Servers corrupting dbfs?
The customer is using something called 2x Application Server.
The dbf is filling up with byte 0 up to 5Gb.
At one incident the file was truncated to 406 bytes instead.
The customer has been forced to restore the file from backup several times in the last months.
There are also users accessing the files directly from their workstations through file sharing.
The file is located on the same computer as the TS users are connecting to, but the TS is using the file share when accessing them so that the access paths are exactly the same. The problem only occurs for the TS users.
The TS users are sometimes experiencing that the application freezes for 30 seconds or so (not processing keystrokes or mouse input) and after that continue like nothing happened.

/Mathias

Check antivirus and backups.
It sure sound to me that something like that is the issue

"Mathias" schreef in bericht
news:29804352.839.1319465732052.JavaMail.geo-discussion-forums@yqnv12...

Is there anyone who's got experience of Terminal Servers corrupting dbfs?
The customer is using something called 2x Application Server.
The dbf is filling up with byte 0 up to 5Gb.
At one incident the file was truncated to 406 bytes instead.
The customer has been forced to restore the file from backup several times
in the last months.
There are also users accessing the files directly from their workstations
through file sharing.
The file is located on the same computer as the TS users are connecting to,
but the TS is using the file share when accessing them so that the access
paths are exactly the same. The problem only occurs for the TS users.
The TS users are sometimes experiencing that the application freezes for 30
seconds or so (not processing keystrokes or mouse input) and after that
continue like nothing happened.

/Mathias

On Mon, 24 Oct 2011 07:15:32 -0700 (PDT), Mathias
<mathias.hakansson@consultec.se> wrote:

Hello Mathias,

>Is there anyone who's got experience of Terminal Servers corrupting dbfs?


Are you using ADS? ADS can give corruption when you mix the use of
axdbfcdx and dbfcdx rdd's.

Dick

mathias,

Are TS and non-TS users working at the same time? It sounds more like
a TCP stack issue based on different OS accessing the shared files
simultaneously. The sympton you describe sounds like the age old
problem of where the RDD tries to write the value with read-back. When
the read-back fails, it writes again... and keeps on doing so until it
succeeds. When this also invovles an append, that is the outcome. Huge
file growth.

Geoff


On Oct 25, 1:15*am, Mathias <mathias.hakans...@consultec.se> wrote:
> Is there anyone who's got experience of Terminal Servers corrupting dbfs?
> The customer is using something called 2x Application Server.
> The dbf is filling up with byte 0 up to 5Gb.
> At one incident the file was truncated to 406 bytes instead.
> The customer has been forced to restore the file from backup several times in the last months.
> There are also users accessing the files directly from their workstationsthrough file sharing.
> The file is located on the same computer as the TS users are connecting to, but the TS is using the file share when accessing them so that the access paths are exactly the same. The problem only occurs for the TS users.
> The TS users are sometimes experiencing that the application freezes for 30 seconds or so (not processing keystrokes or mouse input) and after that continue like nothing happened.
>
> /Mathias

Thanks for your reply Geoff.

They probably are using different OS:es. I think they are using Vista and W7 on the workstations and Server 2003 on the server. No XP machines.
The Application is built with Visual Objects 2.8 SP3. The dbfcdx driver is used. Not ADS.

Do you think this old problem could appear in this environment (which is not "age old", only slightly behind.)

The server is virtualized and runned on the ISP/server hosting company hardware. The file sharing users are connected through a very fast fiber city network, and the TS-users are connecting from another town through an ADSL internet connection.

I think I'll tell them to try tweaking their Anti Virus software, and to investigate what's causing the temporary freezing. Maybe try some of the network settings from below as well.


[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\MRXSmb\Parameters]
"OpLocksDisabled"=dword:00000001

Övriga versioner av windows
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\LanmanWorkstation\Parameters]
"UseOpportunisticLocking"=dword:00000000
"UtilizeNtCaching"=dword:00000000
"UseUnlockBehind"=dword:00000001

/Mathias

On Mon, 24 Oct 2011 22:33:13 -0700 (PDT), Mathias
<mathias.hakansson@consultec.se> wrote:

Hello Mathias,

>Do you think this old problem could appear in this environment (which is not "age old", only slightly behind.)

Mixing different Windows versions shouldn't be the problem; every
client does do that all the times. You could ask them to work a while
with only the same Windows versions on simultaneously to be sure.

Your registry settings reminded me of a long not used .reg file
(W2KOplockParamaters.reg) which I think was needed in Windows 2000
environment. It has a few settings more than you wrote so maybe that
could make the difference?


REGEDIT4

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\MRxSmb\Parameters]
"OpLocksDisabled"=dword:00000001
"CscEnabled"=dword:00000001

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\LanmanServer\Parameters]
"EnableOplocks"=dword:00000000
"CachedOpenLimit"=dword:00000000
"EnableOplockForceClose"=dword:00000001

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\LanmanWorkstation\Parameters]
"UseOpportunisticLocking"=dword:00000000
"UtilizeNtCaching"=dword:00000000
"UseUnlockBehind"=dword:00000001


Dick