This parameter is applicable only to Planning Analytics database running on a Microsoft® Windows 64-bit operating system.
Parameter type: optional, static
If a Planning Analytics database running on a Windows 64-bit operating system is idle for a long period of time, physical memory taken up by the Planning Analytics database will page out to disk. This is a function of the Windows 64-bit operating system and not Planning Analytics database. This can cause performance degradation in large Planning Analytics databases when trying to access data after an idle period.
To maximize performance when running a largePlanning Analytics database on 64-bit Windows, set
LockPagesInMemory=T in the Planning Analytics database configuration. If you change this
parameter value, restart the Planning Analytics database
to apply the new value.
When this parameter is enabled, Windows still trims pages from the Planning Analytics database process space, but does not page them to disk. This benefits Planning Analytics database performance because objects are no longer placed in virtual memory, but instead remain in physical RAM.
When LockPagesInMemory is not present in Planning Analytics database configuration, or if the parameter is set to F, the following behavior is expected:
When a Planning Analytics database running on a Windows 64-bit operating system is idle for a period of time, physical memory taken up by the Planning Analytics database is paged out to disk. This paging to disk happens even if there are no other processes contending for the memory pages. Essentially, Windows leaves the memory pages vacant and available. This is a function of the Windows 64-bit operating system and not Planning Analytics database.
This background paging by the Windows operating system can cause initial performance degradation in large Planning Analytics databases when trying to access Planning Analytics data after an idle period. For example, when the Planning Analytics system has been inactive overnight the first access in the morning will take longer, as the required memory pages containing Planning Analytics data are read from disk back into memory. Also, if the Planning Analytics model is such that there are large cube data areas that are accessed infrequently, the memory holding that cube information may page out to disk. When a request is made for that cube data the request will take longer, as these infrequently used pages must be read back into memory.
LockPagesInMemory=T in Planning Analytics database configuration, the memory pages
containing Planning Analytics data are locked into
memory and are not available for use at any time by any other process. This can make the system
overall perform poorly if there are other tasks that need to run on the Planning Analytics machine. For example, if the machine has 48 GB
of physical memory, and the Planning Analytics database
takes 38 GB to fully load, then there are only 10 GB of physical memory to run any and all other
processes, including system processes. These other processes may perform poorly because they may
force extensive paging activity as the system tries to run them all in what would then essentially
be a 10 GB machine. If a second Planning Analytics
database is started with the same
LockPagesInMemory=T configuration setting, and
that second Planning Analytics database instance would
normally take 12 GB to load, the load will fail since that database can not lock 12 GB of memory, as
there is only 10 GB available.