home | O'Reilly's CD bookshelfs | FreeBSD | Linux | Cisco | Cisco Exam    

UNIX Power Tools

UNIX Power ToolsSearch this book
Previous: 38.6 The Controlling Terminal Chapter 38
Starting, Stopping, and Killing Processes
Next: 38.8 What Are Signals?

38.7 Why ps Prints Some Commands in Parentheses

The reason that some versions of ps , and thus derivatives such as w , sometimes print commands in parentheses [one of our UNIX systems uses square brackets -JP  ]:


ps -f -u jerry

   jerry 29240 29235  0 07:56:19 ttyp1    0:01 sh find_mh_dupes 
   jerry 29259 29240 23 07:57:52 ttyp1    0:07 (egrep)

is that whoever wrote ps liked it that way. The parentheses indicate that the command overwrote its name, or that ps could not find the name, and that ps is printing instead the "accounting name." (The accounting name is the last component of the name given to the exec ( 38.2 ) system call, and is the name used in the system resource usage accounting file.) Basically, ps does this [in the C language -JP  ]:

if (proc->argv == NULL || strcmp(proc->acct_name, proc->argv[0]) != 0)
   printf("(%s)", proc->acct_name);

In the case of a large environment, ps is unable to find the argument vector. This is because it reads only the last few stack pages of each process.

Other versions of ps use completely different mechanisms for locating the command arguments and may never print parentheses.

- CT in net.unix-wizards on Usenet, 13 November 1983

Previous: 38.6 The Controlling Terminal UNIX Power Tools Next: 38.8 What Are Signals?
38.6 The Controlling Terminal Book Index 38.8 What Are Signals?

The UNIX CD Bookshelf Navigation The UNIX CD BookshelfUNIX Power ToolsUNIX in a NutshellLearning the vi Editorsed & awkLearning the Korn ShellLearning the UNIX Operating System