Average Number of Disk I/O Requests in Queue (AverageRequestsInQueue)
This parameter displays the average number of disk I/O requests in the queue and is measured only when the queue is occupied. A high number indicates that system throughput is probably slowing down because of the number of I/O requests for this disk. This can be caused by
- paging and swapping due to low memory (PageFaults)
- overloading an I/O controller's capacity
- having most of the heavily-accessed files on the same disk/controller
Recommendations
Before taking any action, make sure that your system's problem is not paging or swapping because of low memory (PageFaults). You can do many things to help improve disk performance. Most of them require you to be very familiar with the work load and characteristics of your system.
- Balance the I/O load across your disks and controllers. Disk-striping software can help distribute data across multiple disks and/or controllers.
- If you have too many disks on one controller and the load is high for one or more of those disks, you should reduce the number of disks on that controller. Controllers with disks that have a light load can often handle more disks.
- Move heavily accessed files to the fastest disk. You can determine which disk drives are the fasted by comparing their technical specifications.
- Use raw disks rather than file systems for the best performance if the disks are for a heavily-used relational database.
- Add more controllers and/or disk drives.
- Buy disks with faster seek times and raw transfer rates.
Because each AIX system may have vastly different characteristics, no range thresholds have been set by default. With the combination of your experience with your system, your system's work load, and your system's documentation, you can set ranges that are useful.
PATROL properties
Attribute | Default value |
|---|---|
Application class | IAK_Disk |
Command type | Not applicable |
Platform | All platforms |
Icon style | Graph |
Unit | Number (#) |
Border range | Undefined |
Alarm1 range | Undefined |
Alarm2 range | Undefined |
Scheduling (poll time) | Inherits poll time of collector |
Active at installation | Yes |
Parameter type | Consumer |
Value set by |
TrueSight/BMC Helix Operations Management properties
Property | Default value |
|---|---|
Monitor type | Disk |
Key Performance Indicator | No |
Graph by default | No |
Availability | No |
Statistical | Yes |
Config | No |
Response time | No |
Monitor for | No |
Normal distribution | Yes |
Application collection status | No |
Delta | No |
Value format (%d) | No |
This parameter displays the average number of disk I/O requests in the queue and is measured only when the queue is occupied. A high number indicates that system throughput is probably slowing down because of the number of I/O requests for this disk. This can be caused by
- paging and swapping due to low memory (PageFaults)
- overloading an I/O controller's capacity
- having most of the heavily-accessed files on the same disk/controller
Recommendations
Before taking any action, make sure that your system's problem is not paging or swapping because of low memory (PageFaults). You can do many things to help improve disk performance. Most of them require you to be very familiar with the work load and characteristics of your system.
- Balance the I/O load across your disks and controllers. Disk-striping software can help distribute data across multiple disks and/or controllers.
- If you have too many disks on one controller and the load is high for one or more of those disks, you should reduce the number of disks on that controller. Controllers with disks that have a light load can often handle more disks.
- Move heavily accessed files to the fastest disk. You can determine which disk drives are the fasted by comparing their technical specifications.
- Use raw disks rather than file systems for the best performance if the disks are for a heavily-used relational database.
- Add more controllers and/or disk drives.
- Buy disks with faster seek times and raw transfer rates.
Because each AIX system may have vastly different characteristics, no range thresholds have been set by default. With the combination of your experience with your system, your system's work load, and your system's documentation, you can set ranges that are useful.
PATROL properties
Attribute | Default value |
|---|---|
Application class | IAK_Disk |
Command type | Not applicable |
Platform | All platforms |
Icon style | Graph |
Unit | Number (#) |
Border range | Undefined |
Alarm1 range | Undefined |
Alarm2 range | Undefined |
Scheduling (poll time) | Inherits poll time of collector |
Active at installation | Yes |
Parameter type | Consumer |
Value set by |
Central Monitoring Administration properties
Property | Default value |
|---|---|
Monitor type | Disk |
Key Performance Indicator | No |
Graph by default | No |
Availability | No |
Statistical | Yes |
Config | No |
Response time | No |
Monitor for | No |
Normal distribution | Yes |
Application collection status | No |
Delta | No |
Value format (%d) | No |