Managing system performance

Managing system performance

Managing and improving system performance
    Managing CPU resources
        Balancing system work loads
        Detecting runaway processes
        Monitoring user PATH variables
        Identifying heavily loaded system resources
    Managing disk resource usage
        Managing disk space
        Monitoring filesystem use
        Balancing filesystem space: moving user directories
        Controlling directory size
        Selecting a filesystem type
        Selecting a logical block size for a vxfs filesystem
        Selecting a logical block size for a sfs, s5, or ufs filesystem

Monitoring system performance
    Real-time performance monitor (rtpm)
    System monitor
    Disk usage reports
    System performance analysis tools

System activity reporting
    Collecting system activity data automatically
    Collecting system activity data on demand
        Checking file access with sar -a
        Checking buffer activity with sar -b
        Checking system calls with sar -c
        Checking disk activity with sar -d
        Checking page-out and memory freeing activity with sar -g
        Checking kernel memory allocation activity with sar -k
        Checking interprocess communication with sar -m
        Checking page-in activity with sar -p
        Checking queue activity with sar -q
        Checking unused memory with sar -r
        Checking filesystem usage with sar -t
        Checking CPU use with sar -u
        Checking system table status with sar -v
        Checking swapping and switching volume with sar -w
        Checking terminal activity with sar -y
        Checking overall system performance with sar -A
        Checking raw system performance data with sar -R
    Reporting application turnaround with timex

Kernel profiling
    Loading the system profiler (prfld)
    Enabling and disabling the sampling mechanism (prfstat)
    Collecting profiling data (prfdc, prfsnap)
    Formatting the collected data (prfpr)

Displaying and changing parameters

Investigating performance problems
    Outline of typical troubleshooting procedure
    Samples of performance management procedures
        Checking for excess swapping
        Checking for disk slowdowns
        Checking for modem interrupts
        Checking for table overflows
        Analyzing disk resources
        Analyzing memory resource usage
        Shifting the workload to off-hours
    Investigating network troubles

Advanced system performance issues
    Identifying demand loaded DLKMs
    Identifying scheduling problems
    Adjusting filesystem parameters
    Adjusting CPU resource parameters
    Adjusting stack overflow parameters
    Adjusting kernel virtual segments for large mappings
    Configuring the Dedicated Memory feature
    Configuring Dynamically Mapped Shared Memory (DSHM)

Managing large files

Managing Large Physical Memory
    Configuring your system for Large Physical Memory
    Configuring swap space for systems with Large Physical Memory
    Configuring dump space for systems with large physical memory

Quick reference guide to managing performance