AWR指标值计算（一）
众所周知10g引入的AWR比statspack增加了一些OS方面统计信息。AWR报告中有一子项就是OS statistics。今天对这个子项的计算进行了实验。AWR报告的这个信息可以从dba_hist_osstat这个视图查询到。只是我对其中的指标名不是很了解。查了一下文档资料，发现快照的这个信息是从v$osstat取得的。因此查Reference得到如下解释
| Statistic Name | Description |
|---|---|
|
NUM_CPUS |
Number of CPUs or processors available |
|
IDLE_TIME |
Number of hundredths of a second that a processor has been idle, totalled over all processors |
|
BUSY_TIME |
Number of hundredths of a second that a processor has been busy executing user or kernel code, totalled over all processors |
|
USER_TIME |
Number of hundredths of a second that a processor has been busy executing user code, totalled over all processors |
|
SYS_TIME |
Number of hundredths of a second that a processor has been busy executing kernel code, totalled over all processors |
|
IOWAIT_TIME |
Number of hundredths of a second that a processor has been waiting for I/O to complete, totalled over all processors |
|
NICE_TIME |
Number of hundredths of a second that a processor has been busy executing low-priority user code, totalled over all processors |
|
AVG_IDLE_TIME |
Number of hundredths of a second that a processor has been idle, averaged over all processors |
|
AVG_BUSY_TIME |
Number of hundredths of a second that a processor has been busy executing user or kernel code, averaged over all processors |
|
AVG_USER_TIME |
Number of hundredths of a second that a processor has been busy executing user code, averaged over all processors |
|
AVG_SYS_TIME |
Number of hundredths of a second that a processor has been busy executing kernel code, averaged over all processors |
|
AVG_IOWAIT_TIME |
Number of hundredths of a second that a processor has been waiting for I/O to complete, averaged over all processors |
|
AVG_NICE_TIME |
Number of hundredths of a second that a processor has been busy executing low-priority user code, averaged over all processors |
|
OS_CPU_WAIT_TIME |
Total number of hundredths of a second that processes have been in a ready state, waiting to be selected by the operating system scheduler to run |
|
RSRC_MGR_CPU_WAIT_TIME |
Total number of hundredths of a second that Oracle processes have been in a ready state, waiting for CPU to be available for their consumer group in the currently active resource plan |
|
VM_PAGE_IN_BYTES |
Total number of bytes of data that have been paged in due to virtual memory paging |
|
VM_PAGE_OUT_BYTES |
Total number of bytes of data that have been paged out due to virtual memory paging |
|
PHYSICAL_MEMORY_BYTES |
Total number of bytes of physical memory |
|
LOAD |
Current number of processes that are either running or in the ready state, waiting to be selected by the operating-system scheduler to run. On many platforms, this statistic reflects the average load over the past minute. |
|
NUM_CPU_CORES |
Number of CPU cores available (includes subcores of multicore CPUs as well as single-core CPUs) |
|
NUM_CPU_SOCKETS |
Number of CPU sockets available (represents an absolute count of CPU chips on the system, regardless of multithreading or multi-core architectures) |
还算不错，基本上AWR报告上有的都解释了。但是发现报告上有两项分别是NUM_LCPUS和NUM_VCPUS上面的表格没有解释。这是为什么呢？带着这个疑问，又去查官方文档。总算在一个几角旮旯处找到解释：
If Simultaneous Multithreading (SMT) is enabled, the v$osstat view reports 2 additional rows corresponding to the online logical (NUM_LCPUS) and virtual cpus (NUM_VCPUS).
但是这里又产生一个问题，我发现AWR报告对这两个指标的两快照之间的值也进行了差值计算。导致这两项值在报告里面都为0，实际上在两个快照里这两项的值均为4和2。这应该是不合理的，难道是AWR的一个bug？
来自 “ ITPUB博客 ” ，链接：https://blog.itpub.net/7757439/viewspace-446890/ ，如需转载，请注明出处，否则将追究法律责任。
