2.2.2 Example of monitoring memory
By monitoring memory, you can detect a shortage of physical memory or illegal operations by processes.
In Windows, memory consists of a physical memory and a paging file. The following figure shows the concept of memory in Windows.
|
In UNIX, memory consists of a physical memory and a swap file. The following figure shows the concept of memory in UNIX.
|
- Organization of this subsection
(1) Overview of memory monitoring
In general, if there is a memory shortage in the physical memory and paging file (swap file), which use physical areas in RAM, the entire system's performance is affected adversely. However, a memory shortage alone is not always the cause of a bottleneck in the system.
Of the large amount of memory referenced by programs, the areas that are not accessed for more than a specific amount of time are saved in the paging file and loaded into the physical memory as needed. Because the speed of accessing this paging file (swap file) is much lower than the speed of accessing the physical memory, the efficiency of memory utilization is compromised.
Therefore, paging and page faults can be the cause of impaired system processing speed.
-
Paging and swapping refer to the movement of codes and data between the physical memory and the paging file (swap file). Loading data from the paging file (swap file) into the physical memory is called page-in, and saving data from the physical memory into the paging file (swap file) is called page-out.
-
Page faults refers to accessing an area that does not exist in the physical memory.
We recommend that you monitor the efficiency of memory utilization, such as paging and page faults, as well as the memory usage. Note that paging occurs even during normal processing. Measure the baseline during stable system operation to determine an appropriate threshold value.
The following table lists and describes the records and fields that are used for monitoring the memory.
No. |
Record |
Field |
Description of value |
Interpretation of value |
---|---|---|---|---|
1 |
PI |
Paging Pages/sec |
Paging count (per second) |
When the threshold value is exceeded continuously, the memory might be causing a system bottleneck due to frequent paging. Exceeding the threshold value temporarily is acceptable. |
2 |
Page Fault Counts/sec |
Number of page faults (per second) |
When the threshold value is exceeded continuously, the memory might be causing a system bottleneck due to frequent page faults. |
|
3 |
Total Mem Mbytes |
Capacity of physical memory |
Check the capacity of the physical memory. |
|
4 |
Free Mem Mbytes |
Free physical memory capacity |
Check the free physical memory space. |
|
5 |
Used Mem Mbytes |
Amount of physical memory used |
When this value is high, a large amount of physical memory might be in use. |
|
6 |
Used Mem % |
Physical memory usage rate |
When this value is high, a large amount of physical memory might be in use. |
|
7 |
Total Swap Mbytes |
Capacity of virtual memory |
Check the capacity of virtual memory. |
|
8 |
Free Swap Mbytes |
Free virtual memory capacity |
Check the free virtual memory space. |
|
9 |
Used Swap Mbytes |
Amount of virtual memory used |
When the threshold value is exceeded continuously, you might need a larger physical memory. |
|
10 |
Used Swap % |
Virtual memory usage rate |
When the threshold value is exceeded continuously, you might need to expand the paging file. |
|
11 |
Effective Free Mem Mbytes#1 |
Actual free physical memory capacity |
Check the actual free memory space available to the system. |
|
12 |
Effective Used Mem Mbytes#2 |
Amount of actual physical memory used |
When this value is high, a large amount of actual physical memory might be in use. |
- #1
-
- In Linux, AIX, and Solaris
The value is obtained by adding Free Mem Mbytes and the amount of releasable areas assigned to the caches and buffers.
- In Windows and HP-UX
The value is equal to the amount of Free Mem Mbytes.
- #2
-
The value is obtained by subtracting Effective Free Mem Mbytes from Total Mem Mbytes.
A memory shortage can also occur due to defective programs.
Take appropriate action as necessary, such as identifying a process that unnecessarily uses a large amount of memory or whose memory usage increases continuously without limit, or monitor each process's memory usage.
You must use PFM - Agent for Platform to monitor each process's memory usage. For details about how to monitor processes, see the manual JP1/Performance Management - Agent Option for Platform (for Windows systems), or JP1/Performance Management - Agent Option for Platform (for UNIX systems).
(2) Example of a monitoring template for monitoring memory
This subsection describes an example of alarms and reports that are provided as a monitoring template for monitoring memory.
PFM - RM for Platform provides alarms and reports, such as the Available Memory alarm and the Memory Used Status (Multi-Agent) report. To obtain more detailed performance information for the memory, you must monitor various aspects of the memory.
(a) Alarms
The following table lists and describes the memory-related alarms.
No. |
Alarm |
Record |
Field |
Abnormal condition |
Warning condition |
Interpretation of value |
---|---|---|---|---|---|---|
1 |
Available Memory |
PI |
Effective Free Mem Mbytes |
< 3 |
< 4 |
Size of the physical memory actually available to applications below 4 is treated as the warning or abnormal status. When this value becomes smaller than the threshold value set in the warning or abnormal condition, a shortage of physical memory might have occurred. If you find a process that makes excessive use of the memory, you must check the status of the process, and then take appropriate action. If there is no process that is using the memory excessively, you must take an action such as expanding the memory. |
2 |
Page Faults |
Page Fault Counts/sec |
>= 5 |
>= 4 |
A page fault count of 4 or greater per second is treated as the warning or abnormal status. When this value becomes greater than the threshold value set in the warning or abnormal condition, a memory shortage might have occurred. |
|
3 |
Pagescans |
Page Scan Counts/sec |
> 150 |
> 100 |
A page scan count per second that exceeds 100 is treated as the warning or abnormal status. When this value becomes greater than the threshold value set in the warning or abnormal condition, a memory shortage might have occurred. |
|
4 |
Swap Outs |
Swapped-Out Pages/sec |
> 200 |
> 100 |
If more than 100 pages are swapped out per second by swap-out processing, the system treats it as the warning or abnormal status. When this value becomes greater than the threshold value set in the warning or abnormal condition, a memory shortage might have occurred. |
|
5 |
Used Swap Mbytes |
Used Swap Mbytes |
>= 1024#1 |
>= 1024#2 |
If the amount of virtual memory used exceeds the value of the Total Swap Mbytes field, the system treats it as the warning or abnormal status. When this value becomes greater than the threshold value set in the warning or abnormal condition, a memory shortage might have occurred. |
- #1
-
Set a value that is about 90% of the value of the Total Swap Mbytes field.
- #2
-
Set the same value as for the Total Mem Mbytes field.
(b) Reports
The following table lists and describes the memory-related reports.
No. |
Report name |
Information displayed in the report |
---|---|---|
1 |
Memory Used Status (Multi-Agent) |
Displays the status of physical memory usage in multiple systems. |
2 |
Memory Used Status (Multi-Agent) (6.0) |
|
3 |
Memory Used Status |
Displays the status of physical memory usage in the system. |
4 |
Memory Used Status (6.0) |
|
5 |
Pool Nonpaged Status |
Displays the size of physical memory that cannot be paged out. |
6 |
Pool Nonpaged Status (6.0) |
|
7 |
System Overview |
Displays the operation status of the system. |
8 |
System Overview (6.0) |