Using File Flow for Troubleshooting


The previous sections provided you with a lot of detail about the flow of messages through the GroupWise system. This information might seem a little esoteric, though. This section applies the information to some scenarios in which the GroupWise administrator might need to troubleshoot problems with message flow.

Troubleshooting Scenario #1: Can't Send Internet Email

In this scenario, the users in DOMAIN1.PO1 note that they can send email within the post office but can't send email to recipients on the Internet.

Here are the typical troubleshooting steps taken to resolve this problem:

1.

The first thing the administrator does is check the GWIA on DOMAIN3. Obviously, to send to the Internet, users have to be able to send through the GWIA. It is up and running just fine, though. As a test, the administrator sends a quick message to a side myrealbox.com account. It is delivered.

2.

It occurs to the administrator that perhaps the users cannot send to anyone outside their post office. Rather than having one of these users send a test message, the administrator sends a message to one of the users in DOMAIN1.PO1, as well as to a user on DOMAIN2. The logic here is that by the time status messages have returned from DOMAIN2, they should also have returned from DOMAIN1.

3.

Checking the status of the message, the administrator sees that the message to DOMAIN1 remains in a pending state long after the message to DOMAIN2 showed up as being delivered. This means that something is blocking message flow to or from DOMAIN1.

4.

The administrator decides to focus on DOMAIN1.PO1 first (even though the problem might be with other post offices under that domain as well). Opening the PO1\WPCSIN\4 directory in Windows Explorer, the administrator finds nearly 300 files queued up for transport. The oldest file is nearly an hour old.

5.

The administrator checks the message transfer status on the POA console (see Chapter 8, "Configuring the Post Office Agent," for keystrokes, or view the POA through the HTTP interface to identify this), and the POA claims that the connection with the MTA is up and running. So the problem is not with the connection. However, in checking the message transfer status of the POA through the HTTP interface, the administrator notices that of the 300 files queued up for transport, the oldest file is 320MB.

6.

At this point, the administrator has determined the cause of the problem. Someone sent a huge message, and the POA and/or MTA could not handle the file. Even though these agents are multithreaded, all threads must use the same physical connection between MTA and POA, and in this case, the network link did not have the bandwidth to manage the transfer between the post office and the domain. All other messages sitting in the WPCSIN\4 directory are being held up behind this one.

The solution here is not likely to be pretty. The administrator could try to unload the POA to get it to release the file, but it is unlikely that it will unload cleanly. This means that the server will have to be downed and the file cleared out manually. Still, with the application of good message-flow principles, the administrator can resolve the problem without calling technical support.

Lessons from Troubleshooting Scenario #1

There are some basic GroupWise troubleshooting principles you should take from this scenario:

  • The people who report problems are rarely technicians. Don't expect their information to be accurate. Was the information from the people on PO1 accurate? Well, yes, they couldn't send to the Internet, but the way they experienced the problem could lead you to believe that they were having problems only with Internet email. They were, in fact, having problems sending to anyone outside the post office.

  • Focus at the first point of potential failure. In this scenario, that meant focusing on PO1. If users on PO1 are complaining of a problem, log in as a test user on PO1 and witness the problem for yourself.

  • Think message flow. Users say they cannot send to the Internet. What's the first place from PO1 that a message has to go in order to get to the Internet? The POST OFFICE\WPCSIN\X directories, which is therefore the first place to start looking. If there aren't files in the POST OFFICE\WPCSIN\X directory, keep looking upstream until you find the problem.

Troubleshooting Scenario #2: Can't Send Between Post Offices

In this scenario, Bob on PO1 complains that he cannot send to Mary Baker on PO2. He gets the mail message back with an error D101 (user not found). The administrator troubleshoots by first sending to Mary Baker from his or her own mailbox on PO3. Mary gets the email, and no errors are reported. Now the administrator must watch Bob send a message to Mary. Here's what happens:

  1. Bob starts typing Mary B and the name completion fills in the rest.

  2. The administrator has Bob open the address book. In the GroupWise address book, Mary Baker's entry is not grayed out. The administrator has Bob click the Frequent Contacts book. There he points out the entry for Mary Baker that has been grayed out. The administrator has Bob delete this entry from the frequent contacts list.

  3. Now the administrator has Bob send another message to Mary. Name completion executes as usual, and this time the message is delivered normally.

Lessons from Troubleshooting Scenario #2

Here are a couple more GroupWise troubleshooting principles to take home:

  • Watch your users, if you can. The devil is in the details, as they say, and it is not likely that Bob would have reported the details of name completion to his administrator.

  • Isolate the problem. This likely means running a quick test or two. In this scenario, the administrator had to determine whether it was just Bob who couldn't send to Mary, or whether it was everyone.

  • After you've isolated the problem, apply your architectural knowledge. In this case, an address book entry was incorrect, but it was not in the system address book. Knowing that the frequent contacts list is stored in the user database helped the administrator understand how one (and only one) user might get an error sending to one other user.



NOVELL GroupWise 7 Administrator Solutions Guide
Novell GroupWise 7 Administrator Solutions Guide
ISBN: 0672327880
EAN: 2147483647
Year: 2003
Pages: 320
Authors: Tay Kratzer

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net