Re: Is that process running?

[Cameron Simpson <cs zip com au>]

On 10:38 24 Jun 2003, Dan Bar Dov <Dan_Bar_Dov etagon com> wrote:
| This is probably a historical question, yet I'd like to re-open it.
| What is the best way to determine if a process is up?
| 
| Best way = portable across linux distributions, quick, relies on minimum
| availability of resources
| Checkproc for example is used in init.d scripts on SuSE, but is not a
| command in RedHat distribs. A chkproc function is used by RedHat for the
| same purpose, but is not available except for init scripts.

If you know its process id and have the rights to signal it (i.e. you're root
or your-uid == its-uid) then:

	if kill -0 $pid 2>/dev/null
	then  it is up
	else  it is not up
	fi

For most services, a well written daemon will write a pid file (eg
/var/run/daemon-name.pid) from which the pid may be obtained.

| The command pidof(8) is a good candidate. Another is pgrep (a canned ps
| and grep?).
| Pidof is packaged in sysvinit, while pgrep is part of a "ps" package.

I'm not a big fan of pidof etc. Grepping a process listing is inherently
a very dodgy way of locating a process' pid. You've not guarenteed that
no other command uses that name (or one sufficiently similar to match
your grep pattern). If there's more than one match, which do you pick?
Should you pick either? The whole thing makes me think "kludge".

In principle for most things you should be able to obtain a reliable
notion of the pid, usually from a /var/run/daemon.pid file or something
like that.

Another thing you can do is not "check the pid" but "see if the service
is there. Eg examine the output of "netstat -an" for a service listening
on the appropriate port (eg with awk checking for LISTEN and the port
number in the right columns).

Just some thoughts. Cheers,
-- 
Cameron Simpson <cs zip com au> DoD#743
http://www.cskk.ezoshosting.com/cs/

Speed costs money.  How fast do you want to go? - C. Shelby