linux unix性能工具-1 -监测工具简介

§1, 监测工具简介

§1.1 监测工具

可以分为4类:

1,内核统计收集工具

比如: vmstat, mpstat, and netstat.

·         Mpstat: 每个处理器的统计和使用情况

·         Vmstat      内存,运行队列, 统计的CPU使用

·         Iostat: 磁盘I/O子系统操作,带宽和使用

·         Netstat:       网络接口包速率,错误和冲突

·         Kstat:      显示内核统计

·         Sar         系统活动报告

以上命名只有netstat不是动态变化的.

2,进程工具:

比如: prstat, ptree, and pfiles.

·         prstat - report active process statistics

·         ps - report process status

每个进程的信息可由叫ptools或process tools的获取.这些工具位于进程文件系统中,在/proc

·         pargs(1). Display process argument list.

·         pflags(1). Display process flags.

·         pcred(1). Display process credentials.

·         pldd(1). Display process shared object library dependencies.

·         psig(1). Display process signal dispositions.

·         pstack(1). Display process stack.

·         pmap(1). Display process address space mappings.

·         pfiles(1). Display process opened files with names and flags.

·         ptree(1). Display process family tree.

·         ptime(1). Time process execution.

·         pwdx(1). Display process working directory.

Process control is available with various ptools.

·         pgrep(1). Search for a process name string, and return the PID.

·         pkill(1). Send a kill signal or specified signal to a process or process list.

·         pstop(1). Stop a process.

·         prun(1). Start a process that has been stopped.

·         pwait(1). Wait for a process to terminate.

·         preap(1). Reap a zombie (defunct) process.

3,调测工具:

比如: truss and MDB.

·         truss(1). Trace functions and system calls.

·         mdb(1). Debug or control processes.

·         dtrace(1M). Trace, analyze, control, and debug processes.

·         plockstat(1M). Track user-defined locks in processes and threads.

·         cputrack(1). Track per-processor hardware counters for a process.

·         cpustat(1M). Track per-processor hardware counters.

·         busstat(1M). Track interconnect bus hardware counters

4,动态工具;

比如: DTrace

§1.2 Drill-Down 分析

1,监测:

可能使用SunMC, SNMP or sar 以获取全面的信息

2,鉴别

可能使用kstat and procfs tools

3,分析

可能使用TRuss, DTrace, and MDB.

§1.3 其他

本书中,使用率和饱和是有区别的,后者主要是针对队列.

工具的主要来源是:

http://www.solarisinternals.com.

·         Tools bundled with Solaris: based on Kstat, procfs, DTrace, etc.

·         Tools from solarisinternals.com: Memtool and others.

·         Tools from Brendan Gregg: DTraceToolKit and K9Toolkit