3.0 Post Office Agent Problems

For a list of error messages related to the Post Office Agent, see Post Office Agent Error Messages in GroupWise 8 Troubleshooting 1: Error Messages.

This section suggests ways to fix the following problems:

POA Won’t Start

Problem: The POA does not start.
Possible Cause: The /home switch is missing.
Action: Make sure the /home startup switch provides the correct path to the post office directory where the post office database (wphost.db) resides. This switch is required to start the POA and must be provided either in the POA startup file or on the command line when you start the POA. See Using POA Startup Switches in Post Office Agent in the GroupWise 8 Administration Guide.
Possible Cause: The /home switch points to an unavailable location.
Action: Make sure the location specified by the /home startup switch is currently available to the POA. If the post office is located on a different server from where the POA will run, preparations for the connection between the POA and the post office are required:

  • If you are using the NetWare POA, you must use the /dn startup switch or the /user and /password startup switches to enable the POA to log in to the server where the post office is located, or you can provide the username and password in ConsoleOne, on the Post Office Settings page of the Post Office object.

  • If you are using the Linux POA, you must mount the file system where the post office is located to the server where the POA is running.

  • If you are using the Windows POA, you must create the drive mapping to the post office before starting the POA.

See Setting Up the GroupWise Agents in the GroupWise 8 Installation Guide.

Possible Cause: The server where the post office resides is down.
Action: Check the status of the server where the post office resides.
Possible Cause: The POA does not have sufficient rights to the post office directory.
Action: Make sure the network rights in the post office are correct. Start the POA using the /rights switch to determine the specific problem the POA is encountering and make corrections as needed.
Possible Cause: The post office database (wphost.db) is damaged.
Action: If the post office database is available to the POA but cannot be read, it might be damaged. In ConsoleOne, perform database maintenance to correct any problems with the post office database. See Maintaining Domain and Post Office Databases in Databases in the GroupWise 8 Administration Guide.
Possible Cause: The POA server has inadequate resources.
Action: Make sure the server where you are trying to run the POA has adequate resources to run the POA, especially adequate available memory and the current versions of any required network system files. See Agent System Requirements in Installing GroupWise Agents in the GroupWise 8 Installation Guide.

Use your operating system tools to check current server resources.

Possible Cause: The POA is not installed correctly.
Action: Make sure all files required to run the POA are installed. For a list of agent files, see Agent Installation Directories in GroupWise 8 Troubleshooting 3: Message Flow and Directory Structure.
Possible Cause: A remote document storage area is unavailable.
Action: Make sure the remote server where the document storage area is located is running.
Action: Make sure the POA has sufficient rights to log in.
Action: Make sure the /user and /password or /dn switches have been provided in the POA startup file.
Action: As a temporary workaround, you can start the POA with the /noconfig switch, so it can start without trying to access the remote document storage area.
Possible Cause: Language files are missing.
Action: If you are using the /language startup switch to run the POA in a particular language, the corresponding language files must be installed for the POA to run in that language. To determine what language-specific files are required, see Agent Installation Directories in GroupWise 8 Troubleshooting 3: Message Flow and Directory Structure.
Possible Cause: A POA is already running on the server.
Action: If you have defined multiple POAs for the same post office in ConsoleOne, the /name switch is required in each startup file to specify which POA configuration to use when you start each instance of the POA.
Possible Cause: The POA encounters an error condition.
Action: If you receive an error message when trying to start the POA, look it up in Post Office Agent Error Messages in GroupWise 8 Troubleshooting 1: Error Messages.

POA Shuts Down Unexpectedly

Problem: The POA has been running smoothly, but stops unexpectedly.
Action: If the POA agent console is still displayed, exit it. If the normal exit procedure does not work, use the system procedure for terminating a program.

  • If you are using the NetWare POA, use the NetWare unload command. If you have other POAs running on the same server, you should exit them before using the unload command. The unload command unloads all POAs running on the same server and might not enable them to terminate gracefully.

  • If you are using the Linux POA, kill the first POA process. You might need to use kill -9.

  • If you are using the Windows POA, close the POA window.

Action: After the POA agent console is no longer displayed, restart the POA as you normally would. See Setting Up the GroupWise Agents in the GroupWise 8 Installation Guide. If the POA shuts down again, exit it again, reboot the server, then start the POA again.
Action: Set the POA log level to Verbose for troubleshooting. See Using POA Log Files in Post Office Agent in the GroupWise 8 Administration Guide.
Possible Cause: Occasionally, a badly damaged message file can cause the POA to shut down.
Action: Check the contents of the POA input queue in the post office. For the location of the POA input queue in the post office, see Post Office Directory in GroupWise 8 Troubleshooting 3: Message Flow and Directory Structure.

Move the message files out of the input queue subdirectories, start the POA, then copy the message files back in groups, watching the POA carefully to see if it shuts down on a particular message file. If it does, delete the problem message file so normal processing can resume.

Action: Rename the post_office\wpcsout\ofs directory, then start the POA. This creates a new, empty input queue and the POA should run smoothly. Then copy the message files from the 0-7 priority subdirectories of wpcsout\ofs back into the correct priority subdirectory so the POA can process them. Copy them in small groups so the damaged message file can be identified and removed.
Possible Cause: Occasionally, a damaged database in the post office can cause the POA to shut down.
Action: In ConsoleOne, perform maintenance to correct any problems with the databases in the post office. See Maintaining Domain and Post Office Databases and Maintaining User/Resource and Message Databases in Databases in the GroupWise 8 Administration Guide.
Possible Cause: Although increasing the number of POA threads from their default settings can, in many cases, improve POA performance, creating too many POA threads can have undesirable results.
Action: For guidance is setting an appropriate number of POA threads, see Optimizing the POA in Post Office Agent in the GroupWise 8 Administration Guide.
Possible Cause: On a NetWare server, Upgrade Low Priority Threads is set to On.
Action: The POA might run more smoothly if you set Upgrade Low Priority Threads to Off. At the server console where the POA is running, type set upgrade low priority threads=off.
Possible Cause: Another program on the server is interfering with the operation of the POA.
Action: If the POA continues to be unstable, eliminate other programs running on the server. If the POA is stable when another specific program is not running on the same server with it, a conflict might exist between the two programs.

POA Statistics Box Shows Requests Pending

Problem: GroupWise client users are using TCP/IP links to the post office. At the POA console, the Statistics box shows a large number of pending client/server requests.
Action: Increase the number of POA connections so that more users can be serviced by the POA. See Adjusting the Number of Connections for Client/Server Processing in Post Office Agent in the GroupWise 8 Administration Guide.

POA Statistics Box Shows Users Timed Out

Problem: GroupWise client users are using TCP/IP links to the post office. At the POA console, the Statistics box shows a large number of users timed out.
Action: Having users timed out does not indicate a problem with the POA, but rather a problem with users. Users who have timed out are users for which the POA has closed the connection because the GroupWise client was no longer communicating. Timed out users might not be exiting GroupWise normally or might be having other problems with their workstations. The number of timed-out users might tend to increase on a daily basis during the hour after users leave to go home. This is not a problem.

POA Statistics Box Shows Undeliverable Users

Problem: At the POA console, the Statistics box shows a large number of undeliverable users. Undeliverable users can be encountered using either mapped drive connections or TCP/IP connections. Undeliverable users are counted differently from undeliverable messages. For example, a single message could be addressed to 10 users; perhaps 9 users received the message successfully but 1 user was undeliverable.
Possible Cause: If messages cannot be delivered to a particular user, that user might have a damaged user database (userxxx.db).
Action: In ConsoleOne, perform maintenance to correct any problems with the user database so messages can be delivered. See Maintaining User/Resource and Message Databases in Databases in the GroupWise 8 Administration Guide.
Action: Check the POA log file for other reasons why messages cannot be delivered to specific users. See Using POA Log Files in Post Office Agent in the GroupWise 8 Administration Guide.

POA Statistics Box Shows Problem Messages

Problem: At the POA console, the Statistics box shows that problem messages have been encountered.
Action: The problem messages number indicates how many messages could not be processed by the POA. For strategies, see Message Is Dropped in the problem Directory in the Post Office.

POA Redirection List Shows Failed TCP/IP Connection

Problem: At the POA console, client/server statistics show a failed TCP/IP connection.
Action: Under Client/Server Statistics, use Show Redirection List to list existing POAs and the IP addresses of the computers they are running on.
Action: Under Client/Server Statistics, use Check Redirection List to determine which connections are currently valid.
Action: If a connection is listed as failed for a POA, use the ping command to see if the server is alive. If the server does not respond to the PING command, you must resolve the TCP/IP problem before the POA can use the link successfully.
Possible Cause: TCP/IP is not functioning correctly on the POA server.
Action: If the POA is running on a Windows server, make sure TCP/IP is correctly installed and set up on the server where the POA is running.
Possible Cause: Multiple servers are trying to use the same IP address.
Action: Check for conflicting IP addresses between those used by POA servers and those used by other servers. Only one server at a time can use the same IP address.

POA Message Transfer Status Box Shows Closed Link

Problem: The POA is configured to communicate with the MTA by way of TCP/IP. However, the link between the POA and the MTA displays as Closed.
Action: Check the last closure reason. This information can help you determine the source of the problem. Common last closure reasons include:
Action: See also POA Fails to Respond to MTP Configuration Changes in ConsoleOne.

POA Starts Unwanted TCP/IP Thread

Problem: Even though you have started the POA using the /notcpip startup switch or disabled the Enable TCP/IP option in the POA Agent Settings page in ConsoleOne, the POA still starts a TCP/IP thread.
Explanation: When you select Client/Server Only or Client/Server and Direct as the access mode for a post office and use the /notcpip switch when starting a POA, that POA does not accept incoming client/server connections from GroupWise clients. However, it still starts a single TCP/IP handler thread if TCP/IP is configured on the server. The purpose of this TCP/IP thread is to notify any GroupWise clients connecting to another POA in the post office via TCP/IP they should reread the database because a new message has been delivered by a POA that is not using TCP/IP.
Action: To totally disable TCP/IP processing in a post office, set the access mode for the post office to Direct Only and start all POAs servicing that post office with the /notcpip switch, or deselect Enable TCP/IP in the Agent Settings page in ConsoleOne.

POA Fails to Deliver Messages

Problem: The POA is running but no messages are being delivered.
Possible Cause: The post office is closed.
Action: From the MTA console, check the status of the post office from the point of view of the domain. See Displaying MTA Status Information in Monitoring the MTA in the GroupWise 8 Administration Guide. If the post office is closed, messages are not arriving in the post office. Correct the problem with the MTA. See MTA Status Box Shows a Closed Location.
Possible Cause: Message file processing has been turned off.
Action: Make sure that message file processing for the POA has not been turned off using the /nomf, /nomfhigh, or /nomflow switches.

POA Fails to Respond to MTP Configuration Changes in ConsoleOne

Problem: You change a POA link configuration or network address setting in ConsoleOne, but the POA does not respond to the change. For example, you change from a mapped or UNC link to a TCP/IP link between the POA and the MTA, or you move the POA to a different server and change its IP address. If the configuration change does not replicate successfully to the post office database, the MTA link to the post office becomes closed.
Action: If you do not want to stop and restart the POA to open a closed post office link, change the configuration information back to what it was before you tried to change it. The MTA should then be able to open the post office link again. After communication between the POA and MTA is re-established, make the configuration changes again. Wait for the configuration changes to be replicated to the post office database (wphost.db), then start the POA in its new location.
Action: Stop the POA, then start the POA using the /mtpinipaddr and /mtpinport switches to specify the new IP address and port the POA should use for Message Transfer Protocol (MTP) communication with the MTA. After the link is established and all administrative messages have been processed, you do not need to use these startup switches again.
Action: Stop the POA. Rebuild the post office database (wphost.db) to replicate the configuration changes. See Rebuilding Domain or Post Office Databases in Databases in the GroupWise 8 Administration Guide.

POA Fails to Respond to Other Configuration Changes in ConsoleOne

Problem: You change a POA configuration setting in ConsoleOne, but the POA does not respond to the change. For example, you change from direct access to client/server mode, or you want to turn off error mail to the administrator.
Action: Synchronize the post office database (wphost.db) with the domain database (wpdomain.db). See Synchronizing Database Information in Databases in the GroupWise 8 Administration Guide.
Action: Stop the POA. Rebuild the post office database (wphost.db) to replicate the configuration changes. See Rebuilding Domain or Post Office Databases in Databases in the GroupWise 8 Administration Guide.

POA Fails to Update the Post Office Database Version

Problem: You are updating the agent software on a server where both the POA and the MTA are running. The MTA successfully updates the database version for the domain but the POA fails to update the database version for the post office.
Possible Cause: You let the Installation program start both agents automatically. As a result, the POA checked the domain version in the post office database before it had a chance to process the administrative message from the MTA that would update the domain version in the post office database.
Action: Wait until the MTA has updated the domain database. For a large domain database, you might need to wait as much as 20 minutes or more. Verify that the database version has been updated by checking the Domain object’s Identification page in ConsoleOne, then stop and restart the POA to update the post office database. Check the Post Office object’s Identification page in ConsoleOne to verify that the database version has been updated.
Action: If restarting the POA does not update the post office database version:

  1. Compare the dates on the .dc files (gwpo.dc and ngwguard.dc) in the post office directory with the dates on the .dc files in the update source .

  2. If the dates on the .dc files in the post office are older than the dates on the .dc files in the update source, copy the .dc files from the update source into the post office directory.

  3. At the POA console, recover the post office database. See Recovering the Post Office Database Automatically or Immediately in Post Office Agent in the GroupWise 8 Administration Guide.

    When the recovery process is finished, the database version should be updated.

  4. View the Post Office object’s Identification page in ConsoleOne to verify that the database version has been updated.

POA Runs Repeated Recoveries on the Same Database

Problem: The POA continually attempts to recover a particular database but never succeeds.
Action: The recovery operation that the POA can perform while the database is in use is not as powerful as a rebuild operation. Rebuild the problem database. See Maintaining Domain and Post Office Databases and Maintaining User/Resource and Message Databases in Databases in the GroupWise 8 Administration Guide.
Action: If the problem is occurring with a user database and it cannot be rebuilt, re-create the user database. See Re-creating a User Database in Databases in the GroupWise 8 Administration Guide.

POA Starts in the Wrong Language

Problem: You have installed the POA in more than one language and it is starting in a different language than you want.
Action: Start the POA using the /language switch to specify the language.

POA Is Involved with Network Operating System or Hardware Problems

Problem: The POA is interacting with the network operating system or hardware in an undesirable way.
Possible Cause: If you just updated the NetWare POA software, you might not have unloaded the agent engine (gwenn5.nlm), resulting in a series of Loader cannot find public symbol: symbol errors on the server console.
Action: Unload gwenn5.nlm, then start the NetWare POA, so that the newly installed agent engine is loaded along with it.
Possible Cause: The NetWare POA server is running older software, resulting in TCP_HANDLER errors on the server console.
Action: Make sure you are using the most current TCP/IP stack for NetWare.
Possible Cause: The POA server is overburdened, resulting in SYN attacks.
Action: Make sure overall server utilization is not too high. Increase the POA’s TCP/IP resources. See Adjusting the Number of Connections for Client/Server Processing in Post Office Agent in the GroupWise 8 Administration Guide.
Possible Cause: The POA server is overburdened. CPU utilization on the server where the NetWare POA is running jumps to 100% and the POA seems to be taking all available resources. GroupWise and other objects on the server are out of sync with other servers.
Action: Run DSREPAIR.