Project Winproj average percentage of incremental save to full save
Published: June 10, 2010
Applies to: Project Server 2010 | Project Server 2013
Element ID / Rule Name: Project_Winproj_Percentage_Of_Incremental_Save_To_Full_Save
Summary: The first time that a project is created on the Microsoft Project Server application server from Microsoft Project Professional, it is saved using a full save, which means that all of the project data is transmitted and saved in the database used by the Project Server application server. Subsequent saves of the project are saved incrementally, by default. This means that only the project data that has CHANGED is transmitted to the server and saved in the database. This is mainly to optimize performance.
Full saves also occur in the following situations:
If the incremental save fails, a FULL SAVE is attempted.
The "correlated job group" for a particular project is blocked. For example, a project check-in failure can block the correlated job group for that project. After this, when a user attempts to save that same project from Microsoft Project Professional, the "correlated job group" for that project is canceled and a FULL SAVE of a project is performed. For more information about correlated job groups, see the queue documentation.
Note that users are never aware of whether their project is saved using a FULL SAVE or an incremental save. However, the administrator should care about this issue because it directly affects performance.
This alert occurs if the percentage of incremental saves is less than the threshold.
Cause: The number of full saves that are occurring has exceeded the threshold set for this alert.
Possible resolutions include the following:
Use the Queue Management page in Microsoft Project Web App and check for failed jobs to determine which jobs are blocking the correlated job group. Review the Windows NT Event and trace logs to determine whether there are events or traces from the "Project Server Active Cache Save" category that relate to the failed queue jobs that you discovered on the Queue Management page, and troubleshoot accordingly.