Troubleshooting Exchange Startup Problems

Follow

Summary

This Knowledgebase article provides information about troubleshooting Exchange startup problems.

 

More Information

Symptoms

Exchange fails to start:

  • Exchange server not available to the users.
  • The Application Status is 'Failed' in the Neverfail Heartbeat Management Client ( System -> Status & Control tab)

 

Cause

Exchange Services

There are two services that often fail to start on Exchange:

  • System Attendant (SA)
  • Exchange Information Store (IS)

The second depends on the first, so if the SA fails to start, do not expect the IS to start.

SA starting depends on the domain and in particular the visibility of Active Directory (AD). To see AD, servers require correctly configured DNS entries, without which name resolution fails and the server cannot find AD. Typically, forward and reverse lookup addresses are required in DNS.

This document will not cover every aspect of DNS and AD configuration. The steps below detail common problems more advanced configuration issues are not covered here. If the following steps fail then SA will fail to start.

 

Resolution

System Attendant Startup Problems:

  1. Network availability: Verify network availability using the 'ping' command.

    Example: ping 192.168.15.80

    Successful response: "Reply from 192.168.15.80 bytes=32 time=25ms TTL=128"

    Failed response: "Request timed out"

  2. DNS configuration: Verify DNS configuration using the 'nslookup' command.

    Example: nslookup irnbru

 

Successful nslookup:

Server: tr4cl.lab.qa.com
Address: 192.168.15.1

Name: irnbru.lab.qa.com
Address: 192.168.15.80

 

Failed nslookup:

Server: ng-stl-mail.stirling.neverfailgroup.com
Address: 192.168.1.1

*** ng-stl-mail.stirling.neverfailgroup.com can't find irnbru: Non-existent domain

The failed 'nslookup' would suggest that the network adapter is incorrectly configured; it is looking on the wrong subnet for the DNS server. Check the properties for the network card. Is there a gateway to another network? The DNS server should be on the same network as the default route for an adapter.

Use the command 'route print' to check that IP packets are being sent through the correct adapter for each NIC (please refer to the Windows Help, search on ‘route glossary’).

  1. Network

    Extensive network diagnostics can be run with using the 'Netdiag' utility. More verbose tests are available. Use the "netdiag /?" for more options. Please note that you cannot use the netdiag from 2000 Support Tools on a 2003 server.

  2. Active Directory

    Use the ‘Users and Computers’ manager (available on the Exchange Server or Domain Controller via the programs menu) to browse for the Exchange Server to confirm the status of the account.

    Does it have a red cross through the icon? If "Yes", then at some point, the server’s domain account has been disabled. This can happen under the following condition:

      • The computer has been removed from the domain when active directory was not visible.
      • The sever was disabled by right-clicking on the server and selecting disable.
      • If the computer is not present in the 'Users and Computers' browser then this suggests that the computer has been removed from the domain.

    In either case, please refer to Microsoft Knowledge Base article #297295 as there are implications for Exchange. Simply rejoining the domain via 'My Computer' is insufficient.

  3. WMI and LDAP

    WMI and LDAP problems can also contribute to problems starting exchange. However unless the server is in a bad state, then this is unlikely. The following tools are available in the support tools directory on win 2000 / 2003 CDs and perform a comprehensive range of diagnostics.

      • Dcdiag /s:
      • Ldp.exe

    Errors detected in 'dcdiag' may well indicate problems associated with Windows Services. Check that the appropriate windows services are running and look for errors in the event log.

    The ldp.exe tool can be used to connect to the LDAP server running on Active Directory to check that access is possible.

  4. Windows Services

    Exchange services are dependant upon several key Windows services including:

    • Event Log
    • NT LM Security
    • RPC
    • Server
    • Workstation
    • EXIFS (Allows the M: drive to be mapped for user access via Explorer and Web Clients (POP3 and IMAP4) to mailboxes.
    • Although not explicitly defined as a dependency netlogon errors can also cause SA to fail to start.

    Check that these services are all started and set to 'Automatic'. Any errors will be logged in the 'System Event' log. Please refer Microsoft's Knowledgebase for specific errors.

  5. Active Directory Service

    Verify that the following services are running on the AD Server:

    • DNS
    • Netlogon
    • LDAP
    • RPC and RPC Locator services

    Errors reported for DNS will appear in its own event log on the AD Server; the other services will be logged in the System Log.

    If these all work and SA still fails to start, then there is one further course of action. Study the event IDs and Messages logged in the 'Application Event' log and refer to the web sites at the start of the document.

The following tests all rely upon network connectivity. Network connectivity problems can indicate an incorrectly configured Packet Filter Driver. Please refer to the Section on the Neverfail Packet Filter Driver.

  1. Neverfail Packet Filter Driver

    It is possible that the packet filter is in an incorrect state and that the server is actually masked from the network. The following steps can be performed to diagnose whether this is in fact the case:

    1. From a command prompt type: c:\neverfail\r2\bin\nfpktfltr getstate
    2. The response should be "PassThru" on the active server and "Filter" on the passive server.
    3. Check that the correct IP address is being filtered.
    4. From a command prompt type: c:\neverfail\r2\bin\nfpktfltr getfilter
    5. There may be an IP address conflict. Check the 'System' log for IP conflicts. These will cause the adapter’s IP address to be disabled = no domain visibility. Typically, the second adapter to join the network will be disabled.
  2. WAN Considerations.

    Problems are likely to manifest themselves in the following way:

    1. On a switchover, the application fails to start on the Secondary server. This is characterized by:
      • Switchover taking 10 or more mins to complete.
      • The Event Log records an application timeout
    2. Similar problems in 1 switching from the Secondary to the Primary.
    3. Refer to Knowledgebase article #138 - 'How to Perform a Switchback Within the Neverfail Heartbeat for Exchange Environment' for limitations of multiple switchovers.
    4. Refer to Knowledgebase article #342 - 'Testing Routing in a WAN Environment'.

    If you experience problems described in point 1, then check the Start script on the Secondary server because in all likelihood, this will be incorrect.

    1. First, check that the scripts are executing correctly. This can be done by watching the application scripts execute in the 'Application Manager' pane of the Management Client.
    2. The scripts must run as domain administrator so look for credentials being rejected by DNS.
    3. Does the ‘Run As’ user specified in the Application Manager correspond to the domain administrator?
    4. Check that the scripts contain the correct IP addresses for the Primary and Secondary server. The start script on the Primary should be registering the Primary server’s IP address with DNS and removing the Secondary IP address. The Secondary server will be doing the reverse and registering the Secondary server.
  3. Information Store Errors

    The Information Store depends on:

    • The EXIFS service / driver, which is used for reading / writing to the mailboxes via web based clients.
    • System Attendant, see Section System Attendant Startup Problems.
  4. Data corruption

If the information store fails to mount, then you will need to refer to the 'Application Event' log for the reason why. You may find that the logs or databases are corrupt. If the logs or databases are corrupt, use one of the following options:

  • Restore from Backup - If the data is corrupt and a backup exists use it.
  • IsInteg – CAUTION The Isinteg.exe tool can be used to check and to fix the integrity of the information store databases; the private information store Priv.edb; and the public information store Pub1.edb.

    From the command line type:  c:\program files\exchsrvr\bin>isinteg /?

    Refer to Microsoft Knowledge Base article #301460

 

Applies To

Exchange 2000 / 2003

 

Related Information

None

 

KBID-214

0 out of 0 found this helpful

Comments

0 comments

Please sign in to leave a comment.