PC NTMMTA: Err Msg: Drive Is Either Local, Used or Greater...

• Microsoft Mail Multitasking MTA for Windows NT, version 3.5

SYMPTOMS

When you run the Windows NT Multitasking MTA (NT MMTA) External Mail program on either the command line or as a service, the following error may occur in a Monitor window or in the SESSION.LOG when you use Dynamic Drive connections:

   Drive is either local, used or greater than LASTDRIVE
   Initializing data structures. Could not connect to drive

CAUSE

The drive letter used by the External Mail program to dynamically connect to the postoffice is unavailable for use. The drive letter may be in use by one of the following processes:

• A hardware device, such as a drive or CD player.
• The substitute drive assignment is in use.
• The LASTDRIVE statement in the CONFIG.SYS is lower than the drive letter the External Mail program is trying to use.
• The network drive assignment is in use by the Dispatch Mail program (see MORE INFORMATION below).
• You may also get the same error if you run the MS-DOS External.exe from Windows NT and using the Dynadmin table. For the NTMMTA, make sure you are running the correct version of External.exe, located in the Mailexe \External.NT subdirectory.

WORKAROUND

Review the current drives used by the system using the Net Use command at the command prompt to ensure that the device letter selected is not used in the EXTERNAL.INI file. Check any drives connected to local subdirectories with the SUBST command.

Confirm that the LASTDRIVE entry in the CONFIG.SYS is higher than the drive letters used by the External Mail program. Check the DISPATCH.INI file to confirm that Dispatch is not trying to use a drive letter that is assigned to the External Mail program. See the More Information section below for additional information.

STATUS

Microsoft has confirmed this to be a problem in version 3.5 of the Microsoft Mail Multitasking MTA for Windows NT. We are researching this problem and will post new information here in the Microsoft Knowledge Base as it becomes available.

MORE INFORMATION

When you use dynamic drives, and the External Mail program dynamically initializes an instance and uses drive E, then pauses until the next cycle of checking for new mail, it will release its connection to drive E. If Dispatch initializes in order to check the process table and/or run a scheduled Directory Synchronization (Dir-Sync) event, it will also use drive E as the next available drive letter. Dispatch usually releases the drive before the External instance resumes its cycle; however, the External Mail program may try to resume before Dispatch is finished with its process. In this case, the External Mail program will try to use the same dynamic drive Dispatch is currently using (drive E), resulting in the error above.

This can be prevented by initiating all Dispatch instances before initializing the External instances. If a Dispatch service needs to be stopped, ensure that all External instances are stopped, and re-start the Dispatch instance followed by the External instance.

If the NT MMTA is started through batch files or AT commands, you can organize the drive letter assignments to ensure that Dispatch should be assigned the first available drive letters (for example, D, E, F, G, or H). The External instances should be assigned the available drive letters at the end of the drive range (for example, U, V, W, X, Y, or Z). This will reduce the possibility that the drive letters assigned to be used by the External instances for dynamic connections will be selected as the next available drive.

If the NT-MMTA is being run as a service, Dispatch instances will normally start before External instances, but you can still have this problem if a dispatch service is stopped and then restarted manually. You can prevent this by using the Registry Editor (RegEdit) to add all dispatch instances to the DependendOnService parameter of:

   HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\MS Mail External -
   (instance)

   MS Mail Dispatch - (instance)

The Windows NT server must be re-booted for these changes to take effect. You can test the changes after re-booting by:

1. Stop a dispatch instance, and see if all the External instances are automatically stopped.

2. Restart all External instances and make sure the Dispatch instance is automatically restarted when you re-start the first External instance.