Cisco Unified MeetingPlace, Release 7.0 -- How to Resolve Failed Server Issues

From DocWiki

Jump to: navigation, search

Main page: Cisco Unified MeetingPlace, Release 7.0

Up one level: Troubleshooting




Note: When restoring failed servers, users will experience a slight delay before the system regains flawless functionality. This delay is equal to the Load Stats Poll Period value on the Site configuration page * five. The Load Stats Poll Period defaults to one minute. Therefore, the delay defaults to five minutes.


Contents

Web Server Fails During a Conference

Problem: A Web Server fails during a web conference and all users are temporarily disconnected from the web-conferencing portion of their meeting.

Possible Cause: There was an Agent Service problem. Other Web Servers in the cluster detected that a server was down and began to exclude it from the load balancing cluster.

Solution: Tell users to rejoin their web conference by clicking the click-to-attend link in their meeting notification or by accessing the Current Meeting page from the Web. The meeting console client automatically tries to reconnect the user to the server. If this attempt fails, the meeting console attempts to connect to the server designated as the backup for that meeting. If no connection is made after 15 seconds, the client continues attempting to contact the primary, then the backup server. After 30 unsuccessful attempts, the client stops trying and notifies the user that it is unable to reconnect.


Possible Cause: There was a server components problem. Other Web Servers in the cluster were unable to detect that a server was disabled since its Agent Service was fully functional. As a result, the failed Web Server was not excluded from the load balancing cluster and continued to have users routed to it in error.

Solution: In the Windows Services control panel, restart the Cisco Unified MeetingPlace Web Conferencing Service. This should restore web-conferencing functionality.

NOTE: When you restart the Web Server, all manual changes made to the registry are lost.



All Web Servers in a Load Balancing Cluster Have Failed

Problem: All Web Servers in a load balancing cluster are nonfunctional.

Possible Cause: The SQL Server database is on a failed server.

Solution: Restore the database. Users are unable to conduct web conferences until the database is restored.



External Users are Disconnected from a Web Conference

Problem: External users are disconnected from a web conference.

Possible Cause: The Web Server in the DMZ has failed.

Solution: Configure another server as your external Web Server. Users are unable to conduct external meetings until you complete this step. To configure an external Web Server, see Configuring Redirection of External Meetings.



Error Message: Read Failure in Named Pipe in Audio Service

Error Message: Read Failure in Name Pipe in Audio Service.

Explanation: The Cisco Unified MeetingPlace Web Conferencing Service has lost its connection to the Cisco Unified MeetingPlace Audio Service. This could be due to a manual restart or an error.

Recommended Action: Do nothing. The Web Conferencing Service will reconnect automatically; the error message is for information only.



Web Conferencing Does Not Start After Moving Time on the Web Server

Problem: User moved the time backwards then forward on the Web Server. Now Web Conferencing does not start.

Possible Cause: There may be bad "date_begin" values in the pps_acl_quotas table in the MPWEB Slave database used by the Web Server (MPWEB_xxxxx). Those dates are initialized at the time when Web Conferencing is installed so if the time was moved backwards, Web Conferencing will refuse to work because it cannot find the proper "disk quota" permissions.

Solution: Manually change the "date_begin" values pps_acl_quotas table in the MPWEB Slave database to some time in the past relative to the current system time.



Time is Not Synchronized Between the Web Server and the Application Server

Problem: Web Conferencing configures the Windows Time service on the Web Server to synchronize its time with the Application Server. When the Cisco Unified MeetingPlace Web Conferencing master service comes up, it should trigger a time synchronization with the Application Server. This is not happening.

Solution: Complete the following workaround.

  1. Stop the Cisco Unified MeetingPlace Web Conferencing master service.
  2. Look at the time on both the Application Server and the Web Server and try to adjust the clock on the Web Server to match the time on the Application Server within 30 seconds.
  3. Invoke the Windows Time Service to synchronize between the Web Server and the Application Server:
    1. Open a DOS command window.
    2. Enter the following command: w32tm/resync
    3. Enter the following command to see if the time is now within 30 seconds: w32tm /monitor /computer:hostname_AppServer



Cannot Start Cisco Unified MeetingPlace Web Conferencing Services

Problem: A Web Server fails and then does not come back up. Cannot start Cisco Unified MeetingPlace Web Conferencing services.

Possible Cause: You configured shared storage for your Web Server using a Windows account that does not have administrator privileges.

Solution: Reinstall Cisco Unified MeetingPlace Web Conferencing.


Solution: Complete the following steps to manually remove the shared storage entries from the SQL database tables. This will allow you to start Web Conferencing services and reconfigure shared storage using an account that has administrator privileges on the Web Server.

  1. Open SQL Server Enterprise Manager.
    Click Start > All Programs > Microsoft SQL Server > Enterprise Manager.
  2. Navigate to the MPWEB database.
    Click the + signs next to SQL Server Group > LOCAL > Databases > MPWEB to open the appropriate directory trees.
  3. Click Tables in the MPWEB directory.
    A list of tables opens in the right pane.
  4. Right-click Site in the right pane.
  5. Select Open table > Return all rows.
    The Site database table displays.
  6. Scroll to the right and set all entries in the following columns to <NULL>:
    • SvcStartAsDomain
    • SvcStartAsUsername
    • SvcStartAsPassword
    Note: Sometimes the column entries appear to be blank after you enter <NULL>. This is okay.
  7. Navigate to the MPWEB slave database in the left pane of the SQL Server Enterprise Manager.
    Click the + signs next to SQL Server Group > LOCAL > Databases > MPWEB_XXXX_XXXX to open the appropriate directory trees.
  8. Click Tables in the MPWEB_XXXX_XXXX directory.
    A list of tables opens in the right pane.
  9. Right-click PPS_CONFIG in the right pane.
  10. Select Open table > Return all rows.
    The PPS_CONFIG database table displays.
  11. Delete the row that contains the NAME config-shared-storage.


Server Crashes When the Shared Storage Location is Interrupted

Problem: The Web Server crashes when network access to the shared storage location is interrupted.

Solution: Restore access to the shared storage location.


Related Topics

Rating: 5.0/5 (1 vote cast)

Personal tools