I have set up an additional Server to allow for testing. Both servers are running Novell 6.5, have 2GB Ram, 500 GB disk and are running the Universal Driver 3.0.0.2. Only on the test server do I have this problem. I get a Windows error where the Error Signature is: AppName: oinsight.exe AppVer: 9.2.1.0 ModName: lh45clnt.dll

ModVer: 0.0.0.0 Offset: 0000811f

The error seems to be connected to the UD but I don't know where else to check. The UD was copied from the live server and all the files are the same on both servers.

The configuration files (REVPARAM and LH3SRVC.INI) are configured exactly the same except the ServerName is 192.168.0.3 on the live server and 192.168.0.2 on the test server. I tried to change the TCPIPort on the test server to 9988 from 9998 but OI then failed to start. Anyone have a suggestion on how to correct this?

Dave

Dave, The faulting module lh45clnt.dll is the client side portion of the network driver. I've seen this error appear when networking isn't properly configured and two instance of OpenInsight try and lock a record in the same table. Without proper networking in OI 9.x the lock will cause a program fault similar to the message you reported.

Verify that ALL the revparam files on the test server were updated with the correct IP address. If the IP address is different some of the clients could be connecting to the wrong server.

Also, try stopping the LinearHash NLM on the test server and try launching OpenInsight. Does it succeed? It should fail but the test helps prove that the client is connecting to the LinearHash NLM.

Last of all, try changing to the 3.0 Universal client driver. It appears that your oinsight.exe client is using the 4.5 driver. Run the netdrv.exe program to change to the 3.x client driver.

Hi Jared,

This test server has only one user (me) logged in and attempting to do a batch process where one file is opened and records read, updated and written back sequentially. The 'crash' doesn't occur on the same record each time.

I checked all the REVPARAM file on the drive to ensure they were all the same. I stopped LH3 and OI would not start. The error was 'Unable to access Linear Hash files.' 'Error FS231'

I ran netdrv.exe and it was set to 'Universal Driver 3.0.0.2' and I reset it. I still get the 'crash'

Is there a setting somewhere in OI to change from LH45clnt.dll to LH31clnt.dll other than NETDRV.EXE?

The FS231 when the LH3 NLM is stopped is a good sign and expected. It indicates the REVPARAM file is being read and attempting to contact the NLM/Service. Verify that any other revparam files on the test server are the same as the one in the root OpenInsight directory.

Does the Database Manager have tables attached by absolute path or relative path? Is it possible that the application is dynamically attaching tables on the live sever?

After changing to Universal Driver 3.0.0.2 the crash message module is different, correct?

It may not be the same error I have seen in the past, your error could be something different. Novell throws an interesting mix into the problem but based on the message and the fact it is crashing in the client side driver I recommend that you examine other possible places where processes or users may have .LK/.OV files locked.

The Database Manager shows the tables as being in folder F:\AREV\PROPERTY and the contents of the data show to be on the test server. There is no way that there is any data connection to the live server. The workstation attached to the test server is unable to see the live server.

I ran one of the programs to where it 'crashed'. I then checked the server monitor and it show 0 logical record locks. I checked all the LK and OV files that were open and none showed any lock. The server monitor showed only one user (me) logged in.

I don't understand why it 'crashes' on a different record each time.

There did not seem to be any change in the Windows error message. When I first went into NETDRV.EXE, it showed that it was already using Universal Driver 3.0.0.2 and I just reset it.

I may try creating a completely separate network for the file server and one workstation and see if that changes anything.

Your first post says the crash message indicated the fault happened in ModName: lh45clnt.dll. This leads me to believe the 4.5 version client driver was being used. When you changed the client driver with netdrv.exe the ModName should be something similar to lh3clnt.dll. If the mod name isn't changing then something else isn't configured normally.

If you take a clean 9.2.1 installation and copy it to the server can you make it crash using the examples application?

When I first started netdrv.exe, it showed that the driver was Universal Driver 3.0.0.2 and I reset it as Universal Driver 3.0.0.2. After doing that I still got the 'crash' with the ModName: lh45clnt.dll. I don't see any distinction in netdrv.exe between server and client. Am I missing something?

I'm going to try running the live system with the same data that is on the test server over the weekend and see what I get.

View this thread on the Works forum...