Best practices for troubleshooting Project Server queue issues
This article describes best practices to use when troubleshooting Microsoft Office Project Server 2007 queue issues. Some of the recommendations below require an understanding of the Office Project Server 2007 queue system and ULS logging. For more information on these topics, see the links in the "See Also" section.
Use the following best practices when you are troubleshooting queue issues.
Verify that you have applied the latest updates to Project Server 2007. The Infrastructure Update for Microsoft Office Servers applies a number of enhancements to the queue system, including:
Two new filters in the Manage Queue page allow for better problem analysis. The Active jobs filter allows administrators to quickly view all active jobs, which is useful if the system appears to be suffering stress caused by queue jobs. The Blocked jobs filter provides a view of all blocked jobs. It can be used with the Related jobs sub-filter to enable rapid problem diagnosis.
The new Related jobs sub-filter allows administrators to quickly view all jobs related to a specific job. This allows them to quickly see if the job is blocked by a previous job against the same object.
New default selection improvements allow for easier settings configuration. Time intervals show all jobs that have a status of Getting enqueued, Waiting to be processed (all wait types), and Failed and Blocking. Also, Cancel Jobs Getting Enqueued is enabled by default, ensuring that partially queued jobs are visible by default. This is vital for troubleshooting Microsoft Office Project Professional 2007 project-save jobs.
Improved vertical and horizontal displays allow for better us of available screen area and less grid scrolling.
For more details, see Project Server 2007 Infrastructure Update Release (white paper) (http://go.microsoft.com/fwlink/?LinkId=121912). To install the Infrastructure Update for Microsoft Office Servers (32-bit), see the associated download page (http://go.microsoft.com/fwlink/?LinkId=121879&clcid=0x409) on the Microsoft Download Center. To install the Infrastructure Update for Microsoft Office Servers (64-bit), see the associated download page (http://go.microsoft.com/fwlink/?LinkId=121883&clcid=0x409) on the Microsoft Download Center.
When searching for jobs on the Manage Queue page, you may want to expand the Job History date range. The default value for this setting is the current day.
When viewing jobs in the Manage Queue page, be aware that, by default, all jobs that have successfully completed do not appear. You can choose to display all successful jobs by changing the settings in the Job Completion States section of the Manage Queue page.
If you see many jobs in the Manage Queue page that are waiting to be processed, it does not necessarily mean that the queue is not working. Many times it can be caused because there is a job in the correlation to which your jobs belong that is in a "getting queued" or "failed and blocking correlation" state. To test whether your queue is not functioning, you can attempt to send an unrelated job through the queue to see whether it is processed. For example, you can create a resource or custom field that you do not plan to use (you can delete it later). If the test job remains in a "waiting to be processed" state, the Queue service has probably stopped responding and should be restarted. If the test job processes (regardless of whether the result was successful), the problem is likely caused by a failed job that is blocking subsequent jobs in the correlation from processing. In the Manage Queue page, you must locate the job causing the problem. You can filter for jobs for a specific project name and for jobs with a job state of "getting queued" or "failed and blocking correlation". Once you find the problematic job, the appropriate action varies, but most likely cancelling the job would be necessary. Take note of any associated error information in the Error column to help determine what caused the problem.
In the Manage Queue page, if you are cancelling any queue jobs that have "getting queued" status, verify that the Cancel jobs getting enqueued option is selected. This option can be configured in the Advanced Options section of the Manage Queue page. It is turned off by default.
If you are searching for a specific queue job entry in the ULS logs, you can narrow your search by knowing at what time the queue attempted to process the job. (This information is found on the Manage Queue page.) This information allows you to search the ULS log for queue entries with the same timestamp.
When trying to find specific queue job entries in a ULS log, first find the job on the Manage Queue page in Project Web Access. Find the Job Group ID for the job on the Manage Queue page, and search or filter for the value in the ULS logs. This UID value is usually contained in the last column of a queue job entry in the ULS log. You can add the Job Group ID column to the Jobs Grid through the Columns section.
When you are addressing queuing issues in the ULS logs, keep in mind that the more severe issues tend to have a value of High in the Level column. Filter for high levels to address the more severe issues first.