File: queue.html

package info (click to toggle)
queue 1.30.1-4woody2
links: PTS
area: main
in suites: woody
size: 2,388 kB
ctags: 1,630
sloc: ansic: 10,499; cpp: 2,771; sh: 2,640; lex: 104; makefile: 87
file content (1904 lines) | stat: -rw-r--r-- 67,939 bytes
<HTML>
<HEAD>
<!-- Created by texi2html 1.56k from queue.texinfo on 17 May 2000 -->

<TITLE>GNU Queue </TITLE>
</HEAD>
<BODY>
<H1>GNU Queue </H1>
<H2>Load-balancing/batch-processing environment</H2>
<H2>and local rsh replacement</H2>
<H2>@today </H2>
<H2>Version 1.12.9</H2>
<ADDRESS>Werner G. Krebs</ADDRESS>
<P>
<P><HR><P>


<H1><A NAME="SEC1" HREF="queue_toc.html#TOC1">Introduction</A></H1>
<P>
<A NAME="IDX1"></A>
<A NAME="IDX2"></A>
<A NAME="IDX3"></A>


<P>
<A NAME="IDX4"></A>
<A NAME="IDX5"></A>
GNU Queue is a UNIX process network
load-balancing system that features an innovative proxy process
mechanism which allows users to control their remote jobs in a nearly
seamless and transparent fashion. When an interactive remote job is
launched, such as say Matlab, or EMACS interfacing Allegro Lisp, a
proxy process runs on the local end. (You can think of this being
equivalent to a running `telnet' or `rsh' process, but more intelligent.)
By sending signals to the local
proxy - including hitting the suspend key - the process on the remote
end may be controlled. Resuming the proxy process resumes the remote job. The
user's environment is almost completely replicated, including not only
environmental variables, but nice values, rlimits, terminal settings
<A NAME="IDX6"></A>
are all replicated on the remote end. Together with <CODE>MIT_MAGIC_COOKIE_1</CODE>
<A NAME="IDX7"></A>
(or <CODE>xhost +</CODE>) the system is X-windows transparent as well,
<A NAME="IDX8"></A>
provided the users local <CODE>DISPLAY</CODE> variable is set to the fully
qualified pathname of the local machine. 


<P>
One of the most appealing features of the proxy process system even with
experienced users is that asynchronous job control of remote jobs by
the shell is possible and intuitive. One simply runs the stub in the
background under the local shell; the shell notifies the user when the
remote job has a change in status by monitoring the stub daemon. 


<P>
When the remote process has terminated, the proxy process returns the exit
value to the shell; otherwise, the stub simulates a death by the same
signal as that which terminated or suspended the remote job. In this
way, control of the remote process is intuitive even to novice users,
as it is just like controlling a local job from the shell. Many of my
original users had to be reminded that their jobs were, in fact,
running remotely. 


<P>
In addition, Queue also features a more traditional distributed batch
processing environment, with results returned to the user via
email. In addition, traditional batch processing limitations may be
placed on jobs running in either environment (stub or with the email
mechanism) such as suspension of jobs if the system exceeds a certain
load average, limits on CPU time, disk free requirements, limits on
the times in which jobs may run, etc. (These are documented in the
sample <CODE>profile</CODE> file included.)


<P>
Queue may be installed by any user on the system; root privileges are
not required.




<H1><A NAME="SEC2" HREF="queue_toc.html#TOC2">Installing Queue as an Ordinary User</A></H1>
<P>
<A NAME="IDX9"></A>


<P>
Installing GNU Queue as an ordinary user is recommended only if you
<A NAME="IDX10"></A>
<A NAME="IDX11"></A>
<A NAME="IDX12"></A>
lack root (aka, superuser or Unix system administrative privileges) 
on your cluster.


<P>
<A NAME="IDX13"></A>
You do not need to have system administrative privileges to install GNU Queue. 


<P>
<A NAME="IDX14"></A>
However, To allow all users in the cluster to use GNU Queue you should
have your cluster's system administrator install Queue following the 
instructions in the chapter Install By Root. See section <A HREF="queue.html#SEC3">Installation of GNU Queue by System Administrator (Preferred)</A>. 
However, if this is impractical, you may install Queue yourself
without resorting to administrative superuser, or root, privileges by
following the instructions in this chapter.


<P>
<A NAME="IDX15"></A>
<A NAME="IDX16"></A>
<A NAME="IDX17"></A>
<A NAME="IDX18"></A>
<A NAME="IDX19"></A>
Note that, under its default configuration, GNU Queue supports only one installation per
cluster, so if you install GNU Queue as an ordinary user you will be
the only user able to run jobs through it. This can be overcome if another
user edits GNU Queue's header files to change its network port numbers to
avoid a conflict with another copy of GNU Queue running on the same cluster.  


<P>
See section <A HREF="queue.html#SEC10">Installation by Ordinary User</A>, on <CODE>-DHAVE_IDENTD</CODE> and
running an RFC 931 <CODE>identd</CODE> service on cluster when installating GNU Queue as an ordinary user.


<P>
<A NAME="IDX20"></A>
<A NAME="IDX21"></A>
<A NAME="IDX22"></A>
<A NAME="IDX23"></A>
To do this, you will need write access to an NFS directory that
is shared among all hosts in your cluster. In most cases, your
system administrator will have set up your home directory this way.


<P>
Installing GNU Queue for one user: 



<OL>

<LI>

<A NAME="IDX24"></A>
 <A NAME="IDX25"></A>
 
Run <CODE>./configure</CODE> .

When installing as an ordinary user, configure sets the makefile to install GNU Queue into the
current directory. <CODE>queue</CODE> will go in <CODE>./bin</CODE>, <CODE>queued</CODE> daemon will go into <CODE>./sbin</CODE>, <CODE>/com/queue</CODE>
will be the shared spool directory, the host access control list file will go into
<CODE>./share</CODE> and the queued pid files will go into <CODE>./var</CODE> . If you want things to go somewhere else,
run <CODE>./configure --prefix=dir</CODE>, where dir is the top-level directory where you want things to be installed.

<CODE>./configure</CODE> takes a number of additional options that you may wish to be aware of, <CODE>./configure --help</CODE> gives
a full listing of them. <CODE>--bindir</CODE> specifies where <CODE>queue</CODE> goes, <CODE>--sbindir</CODE> specifies where <CODE>queued</CODE> goes,
<CODE>--sharedstatedir</CODE> where the spool directory goes, <CODE>--datadir</CODE> where the host access control file goes,
and <CODE>--localstatefile</CODE> where the <CODE>queued</CODE> pid files go. 

If <CODE>./configure</CODE> fails inelegantly, make sure <CODE>lex</CODE> is installed. GNU <CODE>flex</CODE>
is an implementation of lex available from the FSF, <CODE>http://www.gnu.org</CODE>.

<LI>

Now run <CODE>make</CODE> to compile the programs.

<LI>

If all goes well, <CODE>make install</CODE> will install the programs into directory you specified with <CODE>./configure</CODE>. Missing
directories will be created.

The name of the localhost <CODE>make install</CODE> is being run on will be added to the host access control list if it is not
already there.

<LI>

Try running Queue.

Start up ./queued on the localmachine. (If you did a <CODE>make install</CODE> on the localhost the localhost should already
be in the host access control list file.)

<CODE>./queue --help</CODE> gives a list of options to Queue.

Here are some simple examples:


<PRE>
&#62; queue -i -w -n -- hostname
&#62; queue -i -r -n -- hostname
</PRE>

Here is a more sophisticated example. Try suspending and resuming it with Control-Z and 'fg':


<PRE>
&#62; queue -i -w -p -- emacs -nw
</PRE>

If this example works on the localhost, you want want to add additional hosts to the host
access control list in <CODE>share</CODE> (or <CODE>--datadir</CODE>) and start up queued on these.


<PRE>
&#62; queue -i -w -p -h hostname -- emacs -nw
</PRE>

will run emacs on <CODE>hostname</CODE>. Without the <CODE>-h</CODE> argument, it will run the job on the
<CODE>best</CODE> or <CODE>least-loaded</CODE> host in the ACL. See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>, for details on how
host selection is made.

</OL>

<P>
You can also create additional queues for use with the <CODE>-q</CODE> and <CODE>-d</CODE> commands,
as outlined for root users below. Each spooldir must have a <CODE>profile</CODE> file
associated with it. See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>, for details.




<H1><A NAME="SEC3" HREF="queue_toc.html#TOC3">Installation of GNU Queue by System Administrator (Preferred)</A></H1>
<P>
<A NAME="IDX26"></A>
<A NAME="IDX27"></A>
<A NAME="IDX28"></A>


<P>
<A NAME="IDX29"></A>
If you want to just experiment with Queue on a single host, all you
need is a local directory that is protected to be root-accessible only.
For load-balancing, however,
you will need an NFS directory mounted on all your hosts with
'no_root_squash' (see NFS man pages) option turned on. Unfortunately,
the 'no_root_squash' option is required for load-balancing 
because the file system is
used to communicate information about jobs to be run. The default
spool directory is under the default GNU sharedstatedir, <CODE>/usr/local/com/queue</CODE>.


<P>
no_root_squash option is the GNU/Linux name. The option is named differently
under different platforms, see your NFS man pages for the name of the option
that prevents root mapping to nobody on client requests.


<P>
Installing GNU Queue for cluster-wide usage



<OL>

<LI>

<A NAME="IDX30"></A>
 <A NAME="IDX31"></A>
 

Since non-administrators are generally less sophisticated than system administrators, the
default <CODE>./configure</CODE> option is to install GNU Queue in the local directory for use by
a single user only. 

System administrators need to specify <CODE>--enable-root</CODE> to reconfigure GNU to run with root
privileges. This turns off some options, for example, privileged ports are used instead of
relying on the <CODE>identd</CODE> (RFC 931) service if it is installed. See section <A HREF="queue.html#SEC8">Security Issues</A>,
for a discussion of security issues 

Run <CODE>./configure --enable-root</CODE> .

When installing with the <CODE>--enable-root</CODE> option, configure sets the Makefile to install GNU Queue under
the <CODE>/usr/local</CODE> prefix. <CODE>queue</CODE> will go in <CODE>/usr/local/bin</CODE>, <CODE>queued</CODE> daemon will go into <CODE>/usr/local/sbin</CODE>, 
<CODE>/usr/local/com/queue</CODE> will be the shared spool directory, the host access control list file will go into
<CODE>/usr/local/share</CODE> and the queued pid files will go into <CODE>/usr/local/var</CODE> . If you want things to go somewhere else,
run <CODE>./configure --enable-root --prefix=dir</CODE>, where dir is the top-level directory where you want things to be installed.

<CODE>./configure --enable-root</CODE> takes a number of additional options that you may wish to be aware of, <CODE>./configure --help</CODE> gives
a full listing of them. <CODE>--bindir</CODE> specifies where <CODE>queue</CODE> goes, <CODE>--sbindir</CODE> specifies where <CODE>queued</CODE> goes,
<CODE>--sharedstatedir</CODE> where the spool directory goes, <CODE>--datadir</CODE> where the host access control file goes,
and <CODE>--localstatefile</CODE> where the queued pid files go. 

If <CODE>./configure</CODE> fails inelegantly, make sure lex is installed. GNU Flex
is an implementation of lex available from the FSF, http://www.gnu.org.

<LI>

Now run <CODE>make</CODE> to compile the programs.

<LI>

If all goes well, <CODE>make install</CODE> will install the programs into directory you specified with <CODE>./configure</CODE>. Missing
directories will be created.

The name of the localhost <CODE>make install</CODE> is being run on will be added to the host access control list if it is not
already there.

<LI>

Try running Queue.

Start up ./queued on the localmachine. (If you did a <CODE>make install</CODE> on the localhost the localhost should already
be in the host access control list file.)

<CODE>./queue --help</CODE> gives a list of options to Queue.

Here are some simple examples:


<PRE>
&#62; queue -i -w -n -- hostname
&#62; queue -i -r -n -- hostname
</PRE>

Here is a more sophisticated example. Try suspending and resuming it with Control-Z and 'fg':


<PRE>
&#62; queue -i -w -p -- emacs -nw
</PRE>

If this example works on the localhost, you want want to add additional hosts to the host
access control list in <CODE>share</CODE> (or --datadir) and start up queued on these.


<PRE>
&#62; queue -i -w -p -h hostname -- emacs -nw
</PRE>

will run emacs on <CODE>hostname</CODE>. Without the <CODE>-h</CODE> argument, it will run the job on the
"best" or "least-loaded" host in the ACL. See Profile for details on how
host selection is made. See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>.

</OL>

<P>
You can also create additional queues for use with the <CODE>-q</CODE> and <CODE>-d</CODE> commands,
as outlined for users below. Each spooldir must have a <CODE>profile</CODE> file
associated with it. See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>, for details.




<H1><A NAME="SEC4" HREF="queue_toc.html#TOC4">Setting Up Queue for Cluster-wide Usage</A></H1>
<P>
<A NAME="IDX32"></A>


<P>
<A NAME="IDX33"></A>
<A NAME="IDX34"></A>
The GNU Queue system consists of two components, `queued' which runs
as a daemon on every host in the cluster, and `queue' 
is a user program  that allows users to submit jobs to the system.


<P>
The 'queue' binary contacts queued to learn the relative virtual load
averages (explained in 'profile') on each host, and specifies one on
which to run the job. Queued then forks off a process 
and works together with queue on the local end to control the remote job. 


<P>
<A NAME="IDX35"></A>
Look over the sample 'profile' file, See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>, to learn how to customize batch
queues and load balancing. 'profile' has many options. Among others,
you can configure certain hosts to be submit-only hosts for all or
only certain job classes by turning off job execution in these queues.


<P>
<A NAME="IDX36"></A>
<A NAME="IDX37"></A>
<A NAME="IDX38"></A>


<P>
Add the name of each host in the cluster to the access control list. The default
location for this is either <CODE>share/qhostsfile</CODE> or <CODE>/usr/local/share/qhostsfile</CODE>
depending on how <CODE>./configure</CODE> was invoked.


<P>
Finally, if you are installing GNU Queue cluster-wide,
make sure the spool directory (default is <CODE>/usr/local/com/queue</CODE>) 
is NFS exported root-writable on all systems in
your cluster. In GNU/Linux, this is done by setting the
<CODE>no_root_squash</CODE> option in <CODE>/etc/exports</CODE> (and then running <CODE>/usr/etc/exportfs</CODE>
to cause the system to acknowlege the changes; if /usr/etc/exportfs is
not available on your system, restart <CODE>nfsd</CODE> and the <CODE>portmapper</CODE>.) 


<P>
Other operating system flavors have different names for this option. Read <CODE>nfs(4)</CODE>,
<CODE>exports(4)</CODE> and other <CODE>man</CODE> pages for information on setting the
<CODE>no_root_squash</CODE> equivalent on your operating system flavor.




<H1><A NAME="SEC5" HREF="queue_toc.html#TOC5">Running <CODE>queue</CODE></A></H1>
<P>
<A NAME="IDX39"></A>
<A NAME="IDX40"></A>
<A NAME="IDX41"></A>



<PRE>
queue [-h hostname|-H hostname] [-i|-q] [-d spooldir] 
      [-o|-p|-n] [-w|-r] -- command.options
</PRE>

<P>
<A NAME="IDX42"></A>
<A NAME="IDX43"></A>

<PRE>
qsh  [-l ignored] [-d spooldir] [-o|-p|-n] 
     [-w|-r] hostname command command.options 
</PRE>

<DL COMPACT>

<DT><KBD>-h hostname</KBD>
<DD>
<A NAME="IDX44"></A>
<DT><KBD>--host hostname</KBD>
<DD>
<A NAME="IDX45"></A>
force queue to run on hostname.
<DT><KBD>-H hostname</KBD>
<DD>
<A NAME="IDX46"></A>
<DT><KBD>--robust-host hostname</KBD>
<DD>
<A NAME="IDX47"></A>
Run job on hostname if it is up.
<DT><KBD>-i|-q</KBD>
<DD>
<A NAME="IDX48"></A>
<DT><KBD>--immediate|--queue</KBD>
<DD>
<A NAME="IDX49"></A>
Shorthand for the (<CODE>now</CODE> spooldir) and queue (<CODE>queue</CODE> spooldir).
<DT><KBD>[-d spooldir]</KBD>
<DD>
<A NAME="IDX50"></A>
<DT><KBD>[--spooldir spooldir]</KBD>
<DD>
<A NAME="IDX51"></A>
With <CODE>-q</CODE> option, specifies the name of the batch processing directory, e.g., <CODE>matlab</CODE>
<DT><KBD>-o|-p|-n</KBD>
<DD>
<A NAME="IDX52"></A>
<DT><KBD>--half-pty|--full-pty|--no-pty</KBD>
<DD>
<A NAME="IDX53"></A>
Toggle between half-pty emulation, full-pty emulation (default), and the more efficient no-pty emulation.
<DT><KBD>-w|-r</KBD>
<DD>
<A NAME="IDX54"></A>
<DT><KBD>--wait|--batch</KBD>
<DD>
<A NAME="IDX55"></A>
Toggle between wait (stub daemon; default) and return (mail batch) mode.
<DT><KBD>-v</KBD>
<DD>
<A NAME="IDX56"></A>
<DT><KBD>--version</KBD>
<DD>
<A NAME="IDX57"></A>
Version
<DT><KBD>--help</KBD>
<DD>
<A NAME="IDX58"></A>
List of options
</DL>

<P>
<A NAME="IDX59"></A>
<A NAME="IDX60"></A>
The defaults for <CODE>qsh</CODE> are a slightly different: no-pty emulation is the default, and a hostname argument is
required. A plus (<CODE>+</CODE>) is the wildcard hostname; specifying <CODE>+</CODE> in place of a valid hostname is the
same as not using an <CODE>-h</CODE> or <CODE>-H</CODE> option with <CODE>queue</CODE>. <CODE>qsh</CODE> is envisioned
as a <CODE>rsh</CODE> compatibility mode for use with software that expects a <CODE>rsh</CODE>-like syntax.
This is useful with some MPI implementations; See section <A HREF="queue.html#SEC7">Running GNU Queue with MPI and PVM.</A>.


<P>
<A NAME="IDX61"></A>


<P>
Start the Queue system on every system in
your cluster (as you defined in queue.h) by running <CODE>queued</CODE> or
<CODE>queued -D &#38;</CODE> from the directory in which queued is
installed. 


<P>
The later invocation places queued in debug
mode, with copious error messages and mailings, which is probably a
good idea if you are having problems. Sending queued a <CODE>kill -HUP</CODE> will
force it to re-read the profile files and ACL lists, which is good when you wish to
shut down a queue or add hosts to the cluster. <CODE>queued</CODE> will also periodically
check for modifications to these files.


<P>
If all has gone well at this stage, you may now try submitting a
sample job to the system. I recommend trying something like <CODE>queue -i
-w -p  -- emacs -nw</CODE>. You should be able to background and foreground the
remote EMACS process from the local shell just as if it were running
as a local copy. 


<P>
Another example command is <CODE>queue -i -w -- hostname</CODE> which should
return the best host (i.e., least loaded, as controlled by options in
the profile file; See section <A HREF="queue.html#SEC6">Configure a Job Queue's profile File</A>, to run a job on. 


<P>
The options on queue need to be explained:


<P>
<CODE>-i</CODE> specifies immediate execution mode, placing the job in the <CODE>now</CODE>
spool. This is the default. Alternatively, you may specify either the <CODE>-q</CODE> option, 
which is shorthand for the <CODE>wait</CODE> spool, or use the <CODE>-d
spooldir</CODE> option to place the job under the control of the <CODE>profile</CODE> file
in the <CODE>spooldir</CODE> subdirectory of the spool directory, which must previously
have been created by the Queue administrator.


<P>
In any case, execution of the job will wait until it satisfies the conditions
of the profile file for that particular spool directory, which may
include waiting for a slot to become free. This method of batch processing
is completely compatible with the stub mechanism, although it may
disorient users to use it in this way as they may be unknowingly
forced to wait until a slot on a remote machine becomes available. 


<P>
<CODE>-w</CODE> activates the stub mechanism, which is the default.
The queue stub process will
terminate when the remote process terminates; you may send signals and
suspend/resume the remote process by doing the same to the stub
process. Standard input/output will be that of the 'queue' stub
process. <CODE>-r</CODE> deactivates the stub process; standard input/output will
be via email back to the users; the <CODE>queue</CODE> process will return
immediately. 


<P>
<CODE>-p</CODE> or <CODE>-n</CODE> specifies whether or not a virtual tty should be
allocated at the remote end, or whether the system should merely use
the more efficient socket mechanism. Many interactive processes, such
as <CODE>EMACS</CODE> or <CODE>Matlab</CODE>, require a virtual tty to be present, so the <CODE>-p</CODE>
option is required for these. Other processes, such as a simple
<CODE>hostname</CODE> do not require a <CODE>tty</CODE> and so may be run without the
default <CODE>-p</CODE>. Note that <CODE>queue</CODE> is intelligent and will override
the <CODE>-p</CODE> option if it detects both <CODE>stdio</CODE>/<CODE>stdout</CODE> have been re-directed
to a non-terminal; this feature is useful in facilitating system
administration scripts that allow users to execute jobs. [At some
point we may wish to change the default to <CODE>-p</CODE> as the system
automatically detects when <CODE>-n</CODE> will suffice.] Simple, non-interactive
jobs such as <CODE>hostname</CODE> do not need the less efficient pty/tty
mechanism and so should be run with the <CODE>-n</CODE> option. The <CODE>-n</CODE> option
is the default when <CODE>queue</CODE> is invoked in <CODE>rsh</CODE> compatibility mode
with <CODE>qsh</CODE>.


<P>
The <CODE>--</CODE> with <CODE>queue</CODE> specifies `end of queue options' and everything beyond this
point is interpreted as the command, or arguments to be given to the
command. Consequently, user options (i.e., when invoking queue through
a script front end, may be placed here): 


<P>
<A NAME="IDX62"></A>



<PRE>
#!/bin/sh
exec queue -i -w -p -- sas $*
</PRE>

<P>
or 



<PRE>
#!/bin/sh
exec queue -q -w -p -d sas -- sas $*
</PRE>

<P>
for example. This places queue in immediate mode following
instructions in the <CODE>now</CODE> spool subdirectory (first example) or in
batch-processing mode into the <CODE>sas</CODE> spool subdirectory, provided it
has been created by the administrator. In both cases, stubs are being
used, which will not terminate until the sas process terminates on the
remote end. 


<P>
In both cases, <CODE>pty</CODE>/<CODE>ttys</CODE> will be allocated, unless the user redirects
both the standard input and standard output of the simple invoking
scripts. Invoking queue through these scripts has the additional
advantage that the process name will be that of the script, clarifying
what is the process is. For example, the script might called <CODE>sas</CODE> or
<CODE>sas.remote</CODE>, causing <CODE>queue</CODE> to appear this way in the user's process
list.


<P>
<CODE>queue</CODE> can be used for batch processing by using the <CODE>-q -r -n</CODE>
options, e.g.,



<PRE>
#!/bin/sh
exec queue -q -r -n -d sas -- sas $*
</PRE>

<P>
would run <CODE>SAS</CODE> in batch mode. <CODE>-q</CODE> and <CODE>-d sas</CODE> options force Queue to
follow instructions in the <CODE>sas/profile</CODE> file under Queue's spool
directory and wait for the next available job slot. <CODE>-r</CODE> activates
batch-processing mode, causing Queue to exit immediately and return
results (including stdout and stderr output) via email. 


<P>
The final option, <CODE>-n</CODE>, is the option to disable allocation of a pty on the
remote end; it is unnecessary in this case (as batch mode disables
ptys anyway) but is here to demonstrate how it might be used in a
<CODE>-i -w -n</CODE> or <CODE>-q -w -n</CODE> invocation.




<H1><A NAME="SEC6" HREF="queue_toc.html#TOC6">Configure a Job Queue's profile File</A></H1>

<P>
<A NAME="IDX63"></A>
<A NAME="IDX64"></A>
<A NAME="IDX65"></A>
<A NAME="IDX66"></A>
<A NAME="IDX67"></A>
<A NAME="IDX68"></A>
<A NAME="IDX69"></A>
<A NAME="IDX70"></A>
<A NAME="IDX71"></A>
<A NAME="IDX72"></A>
<A NAME="IDX73"></A>
<A NAME="IDX74"></A>


<P>
<A NAME="IDX75"></A>
Under <CODE>/usr/spool/queue</CODE> you may create several directories
for batch jobs, each identified with the class of the
batch job (e.g., <CODE>sas</CODE> or <CODE>splus</CODE>). You may then place
restrictions on that class, such as maximum number of
jobs running, or total CPU time, by placing a <CODE>profile</CODE>
file like this one in that directory.


<P>
However, the <CODE>now</CODE> queue is mandatory; it is the 
directory used by the <CODE>-i</CODE> mode (immediate moe)
of queue to launch jobs over the network
immediately rather than as batch jobs.


<P>
Specify that this queue is turned on:


<P>
<A NAME="IDX76"></A>

<PRE>
exec on
</PRE>

<P>
The next two lines in <CODE>profile</CODE> may be set to an email address
rather than a file; the leading <CODE>/</CODE> identifies
then as file logs. Files now beginning with <CODE>cf</CODE>,<CODE>of</CODE>, or <CODE>ef</CODE> are ignored
by the queued:


<P>
<A NAME="IDX77"></A>
<A NAME="IDX78"></A>

<PRE>
mail /usr/local/com/queue/now/mail_log
supervisor /usr/local/com/queue/now/mail_log2
</PRE>

<P>
Note that <CODE>/usr/local/com/queue</CODE> is our spool directory, and <CODE>now</CODE> is
the job batch directory for the special <CODE>now</CODE> queue (run via the <CODE>-i</CODE>
or immediate-mode flag to the queue executable), so these files
may reside in the job batch directories.


<P>
<A NAME="IDX79"></A>
<A NAME="IDX80"></A>
The <CODE>pfactor</CODE> command is used to control the likelihood
of a job being executed on a given machine. Typically, this is done
in conjunction with the <CODE>host</CODE> command, which specifies that the option
on the rest of the line be honored on that host only.


<P>
In this example, <CODE>pfactor</CODE> is set to the relative MIPS of each
machine, for example:



<PRE>
host fast_host pfactor 100
host slow_host pfactor  50
</PRE>

<P>
Where <CODE>fast_host</CODE> and <CODE>slow_host</CODE> are the hostnames of the respective machines.


<P>
<A NAME="IDX81"></A>
<A NAME="IDX82"></A>
<A NAME="IDX83"></A>
This is useful for controlling load balancing. Each
queue on each machine reports back an `apparant load average'
calculated as follows:


<P>
1-min load average/ (( max(0, vmaxexec - maxexec) + 1)*pfactor)


<P>
The machine with the lowest apparant load average for that queue
is the one most likely to get the job.


<P>
Consequently, a more powerful <CODE>pfactor</CODE> proportionally reduces the load average
that is reported back for this queue, indicating a more 
powerful system. 


<P>
<A NAME="IDX84"></A>
<A NAME="IDX85"></A>
Vmaxexec is the "apparant maximum" number of jobs allowed to execute in
this queue, or simply equal to maxexec if it was not set.
The default value of these variables is large value treated
by the system as infinity.



<PRE>
host fast_host vmaxexec 2
host slow_host vmaxexec 1
maxexec 3
</PRE>

<P>
The purpose of <CODE>vmaxexec</CODE> is to make the system appear fully loaded
at some point before the maximum number of jobs are already
running, so that the likelihood of the machine being used
tapers off sharply after <CODE>vmaxexec</CODE> slots are filled.


<P>
Below <CODE>vmaxexec</CODE> jobs, the system aggressively discriminates against
hosts already running jobs in this Queue.


<P>
In job queues running above <CODE>vmaxexec</CODE> jobs, hosts appear more equal to the system, 
and only the load average and <CODE>pfactor</CODE> is used to assign jobs. The theory here is that above <CODE>vmaxexec</CODE> jobs, the hosts are fully saturated, and the load average is a better indicator than the simple number of jobs running in a job queue of where
to send the next job.


<P>
Thus, under lightly-loaded situations, the system routes jobs around hosts 
already running jobs in this job queue. In more heavily loaded situations,
load-averages and <CODE>pfactor</CODE>s are used in determining where to run jobs. 


<P>
Additional options in <CODE>profile</CODE>


<DL COMPACT>

<DT><CODE>exec</CODE>
<DD>
<A NAME="IDX86"></A>
on, off, or drain. Drain drains running jobs.

<DT><CODE>minfree</CODE>
<DD>
<A NAME="IDX87"></A>
disk space on specified device must be at least this free.

<DT><CODE>maxfree</CODE>
<DD>
<A NAME="IDX88"></A>
maximum number of jobs allowed to run in this queue.

<DT><CODE>loadsched</CODE>
<DD>
<A NAME="IDX89"></A>
1 minute load average must be below this value to launch new jobs.

<DT><CODE>loadstop</CODE>
<DD>
<A NAME="IDX90"></A>
if 1 minute load average exceeds this, jobs in this queue are suspended until it drops again.

<DT><CODE>timesched</CODE>
<DD>
<A NAME="IDX91"></A>
Jobs are only scheduled during these times

<DT><CODE>timestop</CODE>
<DD>
<A NAME="IDX92"></A>
Jobs running will be suspended outside of these times

<DT><CODE>nice</CODE>
<DD>
<A NAME="IDX93"></A>
Running jobs are at least at this nice value

<DT><CODE>rlimitcpu</CODE>
<DD>
<A NAME="IDX94"></A>
maximum cpu time by a job in this queue

<DT><CODE>rlimitdata</CODE>
<DD>
<A NAME="IDX95"></A>
maximum data memory size by a job

<DT><CODE>rlimitstack</CODE>
<DD>
<A NAME="IDX96"></A>
maximum stack size

<DT><CODE>rlimitfsize</CODE>
<DD>
<A NAME="IDX97"></A>
maximum fsize

<DT><CODE>rlimitrss</CODE>
<DD>
<A NAME="IDX98"></A>
maximum resident portion size.

<DT><CODE>rlimitcore</CODE>
<DD>
<A NAME="IDX99"></A>
maximum size of core dump

</DL>

<P>
These options, if present, will only override the
user's values (via queue) for these limits if they are lower
than what the user has set (or larger in the case of <CODE>nice</CODE>).




<H1><A NAME="SEC7" HREF="queue_toc.html#TOC7">Running GNU Queue with MPI and PVM.</A></H1>

<P>
<A NAME="IDX100"></A>
<A NAME="IDX101"></A>
Many MPI implementations (such as the free MPICH implementation) allow you to
specify a replacment utility for rsh/remsh to propagate processes.


<P>
<A NAME="IDX102"></A>
<A NAME="IDX103"></A>
Just use <CODE>qsh</CODE> as the replacement. Be sure the <CODE>QHOSTSFILE</CODE> lists all hosts
known to the MPI implementation, and the queued is running on them.


<P>
You have three options: place a <CODE>+</CODE> in the MPI hosts file for each job-slot
you want MPI to be able to start, explicitly list Queue's hosts in the 
MPI host file, or use a combination of <CODE>+</CODE> wild-cards and explicitly listed
hosts in MPI's host file.


<P>
<A NAME="IDX104"></A>
<A NAME="IDX105"></A>
<A NAME="IDX106"></A>
The <CODE>+</CODE> is GNU Queue's wild-card character for the hostname when it is invoked
using <CODE>qsh</CODE>. It simply means that Queue should decide what host the process should
run on, which is the default behavior for Queue. Specifying a host instead of
using the <CODE>+</CODE> with qsh is equivalent to the <CODE>-h</CODE> option with the regular queue 
command-line syntax.


<P>
<A NAME="IDX107"></A>
By placing <CODE>+</CODE>s in the MPI host file, MPI will pass <CODE>+</CODE> as the name of the
host for that job slot to GNU Queue, which, in turn, will decide where the job
should actually run.


<P>
<A NAME="IDX108"></A>
By running jobs through GNU Queue this way, GNU Queue becomes aware of jobs
submitted by MPI, and can route non-MPI jobs around them. Normally, you would
want to use a job queue (<CODE>-j</CODE> option) which has a low <CODE>vmaxexec</CODE> set and a high
maxexec, so that MPI's jobs will continue to run, but GNU Queue will
aggressively try to route jobs to other hosts the moment the job queue
begins the fill.


<P>
<A NAME="IDX109"></A>
<A NAME="IDX110"></A>
GNU Queue's load scheduling algorithm is smarter than that of many MPI
implementations, which frequently treat all hosts as equal and implement
a round-robin algorithm for deciding which to host to run a job on. GNU
Queue, on the other hand, can take load-averages, CPU power differences (via
<CODE>profile</CODE> file specifiers), and other factors into account when deciding
on which host to send a particular job to.  


<P>
<A NAME="IDX111"></A>
<CODE>qsh</CODE> represent a stage-1 hook for MPI. Our development team (See section <A HREF="queue.html#SEC13">Getting Help</A>,
for information on joining the development team) is currently working on a stage-2 hook,
in which MPI becomes aware of GNU Queue jobs as well, allowing them to work as an
integrated scheduling team.


<P>
<A NAME="IDX112"></A>
Support for PVM is currently in development as well.




<H1><A NAME="SEC8" HREF="queue_toc.html#TOC8">Security Issues</A></H1>



<H2><A NAME="SEC9" HREF="queue_toc.html#TOC9">Installation by System Administrator</A></H2>
<P>
<A NAME="IDX113"></A>


<P>
Security is always a concern when granting root privileges to software.


<P>
I was security conscious and knowledgeable about UNIX security issues
when I wrote queue. It should be paranoid in all the right places, at
least provided that the spool directory is root-accessible only (standard
installation) or user-accessible (installation by ordinary user)
only.


<P>
Critical ports allow connections only by hosts in the access control list.
Standard checks (TCP/IP wrapper-style) are made to prevent DNS spoofing
and IP forwarding as much as possible. In addition, connections must
be made from privileged ports (root installation version). <CODE>queue.c</CODE> and <CODE>queued.c</CODE>
run with least-privileges, revoking root privileges as soon as they have
verified information and acquired a privileged port.


<P>
Moreover, at the time of this writing the source code has been available
for a number of months and has been used at numerous installations, including
some concerned with security.


<P>
However, this does not guarantee that security holes do no exist.
It is important that security-conscious users scrutinize the source code and 
report any potential security problems to <CODE>bug-queue@gnu.org</CODE>. By promptly
reporting security issues you will be supporting free software by
ensuring that the public availability of source code is a security
asset.




<H2><A NAME="SEC10" HREF="queue_toc.html#TOC10">Installation by Ordinary User</A></H2>
<P>
<A NAME="IDX114"></A>


<P>
In this installation mode, GNU Queue
takes many of the same precautions for these users as when it has been
installed cluster-wide by a system administrator.


<P>
Unfortunately, when Queue is installed by an ordinary user, privileged
ports are not available. This might make it possible for a malicious user already having a shell
account on the same cluster to have <CODE>queued</CODE> or <CODE>queue</CODE> try to spoof each other.


<P>
To close this hole, Queue uses the one-way function <CODE>crypt(3)</CODE> and a cookie passed over NFS to allow <CODE>queued</CODE> and
<CODE>queue</CODE> to authenticate each other. These cookies are used in the root version
as well to prevent port confusion by <CODE>queued</CODE> trying to connect to a <CODE>queue</CODE>
that has earlier died, although they aren't useful from a security standpoint.


<P>
<A NAME="IDX115"></A>
<A NAME="IDX116"></A>
<A NAME="IDX117"></A>
<A NAME="IDX118"></A>
<A NAME="IDX119"></A>
When GNU Queue is compiled with the <CODE>-DHAVE_IDENTD</CODE> 
(and <CODE>-DNO_ROOT</CODE>),
<CODE>queued</CODE> and <CODE>queue</CODE> also use the <CODE>identd</CODE> service (RFC 931) 
to prevent spoofing by checking the ownership of remote sockets
within the cluster. For this to work proplery, <CODE>identd</CODE> must be running on all your cluster hosts, return accurate information (either the
user's login name as given in the password file or his/her uid), and at 
least accept connections from within the cluster in a reasonable amount of
time. The <CODE>./configure</CODE> script
tries to set <CODE>-DHAVE_IDENTD</CODE> automatically based on whether or not your host accepts local
connections to port 113, but some systems intentionally allow identd to output
bogus information for privacy reasons, and <CODE>-DHAVE_IDENTD</CODE>
should not be set on these; if this is the case, you may need to re-compile GNU Queue with
<CODE>HAVE_IDENTD</CODE> undefined in <CODE>config.h</CODE>. Fortunately,
<CODE>queue</CODE> will normally complain immediately if <CODE>-DHAVE_IDENTD</CODE>
is set when it shouldn't be.


<P>
To get around the performance hit of calling the <CODE>crypt(3)</CODE>, one-way
functions are not used if spoofing queued is impossible due to
privilege ports (root installation) or 
authenticated ports (<CODE>HAVE_IDENTD</CODE> service), so running <CODE>identd</CODE>
with GNU Queue or installing GNU Queue cluster-wide as root may offer
a slight performance advantange. In sites which normally send user 
passwords over the network in cleartext it is not expected to substancial
improve secure over the cookie passing mechanism, however.


<P>
These cookies are passed in plaintext, which means that a malicious user 
might be able to observe the NFS network traffic between the hosts and, having shell access on the cluster, might still be able to spoof <CODE>queue</CODE> or 
<CODE>queued</CODE>. Since most sites send UNIX account passwords  
over the network in cleartext as well, this is only of concern
in very secure sites that do not pass passwords in cleartext over the network.


<P>
<A NAME="IDX120"></A>
In the rare event that your site is a very secure site that does not send
passwords in cleartext, and you are compiling Queue without root privileges,
you should have your administrator install the <CODE>identd</CODE> (RFC 931)
service and re-run <CODE>./configure</CODE> to ensure <CODE>HAVE_IDENTD</CODE> is defined
in <CODE>config.h</CODE>.
<A NAME="IDX121"></A>
<A NAME="IDX122"></A>
<A NAME="IDX123"></A>
If your very secure site prefers to spoof <CODE>identd</CODE> for privacy reasons, 
your administrator may be able to restrict <CODE>identd</CODE> access
with <CODE>tcp_wrapper</CODE> or install an accurate <CODE>identd</CODE> on a
non-standard port which could restrict connections to within the cluster
via <CODE>tcp_wrapper</CODE>. You would need to re-compile GNU Queue with this 
new port number set in <CODE>ident.c</CODE>. Another option is to have your 
system administrator install Queue cluster-wide; this uses privileged ports
and therefore may operate securely with resorting to <CODE>identd</CODE>.


<P>
These concerns do not apply when Queue has been installed cluster-wide by
root (<CODE>NO_ROOT</CODE> is not defined), 
because privileged ports are then available.




<H1><A NAME="SEC11" HREF="queue_toc.html#TOC11">Sending Feedback and Getting Help</A></H1>



<H2><A NAME="SEC12" HREF="queue_toc.html#TOC12">Feedback</A></H2>
<P>
<A NAME="IDX124"></A>
<A NAME="IDX125"></A>
<A NAME="IDX126"></A>
<A NAME="IDX127"></A>


<P>
PLEASE SEND US FEEDBACK ON QUEUE!


<P>
Whether you have a queue-tip, are queue-less about how to solve a
problem, or simply have another bad queue joke, que[ue] us in at
<CODE>bug-queue@gnu.org</CODE> and we'll take our que[ue] from you on how best
to improve the software and documentation.




<H2><A NAME="SEC13" HREF="queue_toc.html#TOC13">Getting Help</A></H2>
<P>
<A NAME="IDX128"></A>
<A NAME="IDX129"></A>


<P>
The application's homepage is
<CODE>http://queue.sourceforge.net</CODE>.


<P>
Bug reports should be sent to the bug list <CODE>bug-queue@gnu.org</CODE>. 


<P>
<A NAME="IDX130"></A>
Users are encouraged to subscribe to and request assistance from the 
development list, `queue-developers', as well.


<P>
<A NAME="IDX131"></A>
At the time of this writing, the list was working on several fun 
projects, including improved MPI
&#38; PVM support, secure socket connections, AFS &#38; Kerberos support. We're also
porting and improving a nifty utility that lets you monitor and control 
the execution of Queue jobs throughout your cluster. The list is a great way 
to tap into the group's expertise and keep up with the latest developments.


<P>
So, come join the fun and keep up with the latest developments by visiting
<CODE>http://lists.sourceforge.net/mailman/listinfo/queue-developers</CODE>.


<P>
It is also possible to subscribe from the application's homepage.
 
It is <CODE>http://queue.sourceforge.net</CODE>.


<P>
<A NAME="IDX132"></A>
<A NAME="IDX133"></A>
<A NAME="IDX134"></A>
At the time of this writing, GNU Queue is being maintained by
GNU Queue's primary author, Werner G. Krebs.




<H1><A NAME="SEC14" HREF="queue_toc.html#TOC14">GNU GENERAL PUBLIC LICENSE</A></H1>


<PRE>
Copyright (C) 1989, 1991 Free Software Foundation, Inc.
675 Mass Ave, Cambridge, MA 02139, USA

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
</PRE>



<H2><A NAME="SEC15" HREF="queue_toc.html#TOC15">Preamble</A></H2>

<P>
  The licenses for most software are designed to take away your
freedom to share and change it.  By contrast, the GNU General Public
License is intended to guarantee your freedom to share and change free
software--to make sure the software is free for all its users.  This
General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors commit to
using it.  (Some other Free Software Foundation software is covered by
the GNU Library General Public License instead.)  You can apply it to
your programs, too.


<P>
  When we speak of free software, we are referring to freedom, not
price.  Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
this service if you wish), that you receive source code or can get it
if you want it, that you can change the software or use pieces of it
in new free programs; and that you know you can do these things.


<P>
  To protect your rights, we need to make restrictions that forbid
anyone to deny you these rights or to ask you to surrender the rights.
These restrictions translate to certain responsibilities for you if you
distribute copies of the software, or if you modify it.


<P>
  For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the recipients all the rights that
you have.  You must make sure that they, too, receive or can get the
source code.  And you must show them these terms so they know their
rights.


<P>
  We protect your rights with two steps: (1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software.


<P>
  Also, for each author's protection and ours, we want to make certain
that everyone understands that there is no warranty for this free
software.  If the software is modified by someone else and passed on, we
want its recipients to know that what they have is not the original, so
that any problems introduced by others will not reflect on the original
authors' reputations.


<P>
  Finally, any free program is threatened constantly by software
patents.  We wish to avoid the danger that redistributors of a free
program will individually obtain patent licenses, in effect making the
program proprietary.  To prevent this, we have made it clear that any
patent must be licensed for everyone's free use or not licensed at all.


<P>
  The precise terms and conditions for copying, distribution and
modification follow.




<H2><A NAME="SEC16" HREF="queue_toc.html#TOC16">TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION</A></H2>


<OL>
<LI>

This License applies to any program or other work which contains
a notice placed by the copyright holder saying it may be distributed
under the terms of this General Public License.  The "Program", below,
refers to any such program or work, and a "work based on the Program"
means either the Program or any derivative work under copyright law:
that is to say, a work containing the Program or a portion of it,
either verbatim or with modifications and/or translated into another
language.  (Hereinafter, translation is included without limitation in
the term "modification".)  Each licensee is addressed as "you".

Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.  The act of
running the Program is not restricted, and the output from the Program
is covered only if its contents constitute a work based on the
Program (independent of having been made by running the Program).
Whether that is true depends on what the Program does.

<LI>

You may copy and distribute verbatim copies of the Program's
source code as you receive it, in any medium, provided that you
conspicuously and appropriately publish on each copy an appropriate
copyright notice and disclaimer of warranty; keep intact all the
notices that refer to this License and to the absence of any warranty;
and give any other recipients of the Program a copy of this License
along with the Program.

You may charge a fee for the physical act of transferring a copy, and
you may at your option offer warranty protection in exchange for a fee.

<LI>

You may modify your copy or copies of the Program or any portion
of it, thus forming a work based on the Program, and copy and
distribute such modifications or work under the terms of Section 1
above, provided that you also meet all of these conditions:


<OL>
<LI>

You must cause the modified files to carry prominent notices
stating that you changed the files and the date of any change.

<LI>

You must cause any work that you distribute or publish, that in
whole or in part contains or is derived from the Program or any
part thereof, to be licensed as a whole at no charge to all third
parties under the terms of this License.

<LI>

If the modified program normally reads commands interactively
when run, you must cause it, when started running for such
interactive use in the most ordinary way, to print or display an
announcement including an appropriate copyright notice and a
notice that there is no warranty (or else, saying that you provide
a warranty) and that users may redistribute the program under
these conditions, and telling the user how to view a copy of this
License.  (Exception: if the Program itself is interactive but
does not normally print such an announcement, your work based on
the Program is not required to print an announcement.)
</OL>

These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.

<LI>

You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:


<OL>
<LI>

Accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of Sections
1 and 2 above on a medium customarily used for software interchange; or,

<LI>

Accompany it with a written offer, valid for at least three
years, to give any third party, for a charge no more than your
cost of physically performing source distribution, a complete
machine-readable copy of the corresponding source code, to be
distributed under the terms of Sections 1 and 2 above on a medium
customarily used for software interchange; or,

<LI>

Accompany it with the information you received as to the offer
to distribute corresponding source code.  (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form with such
an offer, in accord with Subsection b above.)
</OL>

The source code for a work means the preferred form of the work for
making modifications to it.  For an executable work, complete source
code means all the source code for all modules it contains, plus any
associated interface definition files, plus the scripts used to
control compilation and installation of the executable.  However, as a
special exception, the source code distributed need not include
anything that is normally distributed (in either source or binary
form) with the major components (compiler, kernel, and so on) of the
operating system on which the executable runs, unless that component
itself accompanies the executable.

If distribution of executable or object code is made by offering
access to copy from a designated place, then offering equivalent
access to copy the source code from the same place counts as
distribution of the source code, even though third parties are not
compelled to copy the source along with the object code.

<LI>

You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License.  Any attempt
otherwise to copy, modify, sublicense or distribute the Program is
void, and will automatically terminate your rights under this License.
However, parties who have received copies, or rights, from you under
this License will not have their licenses terminated so long as such
parties remain in full compliance.

<LI>

You are not required to accept this License, since you have not
signed it.  However, nothing else grants you permission to modify or
distribute the Program or its derivative works.  These actions are
prohibited by law if you do not accept this License.  Therefore, by
modifying or distributing the Program (or any work based on the
Program), you indicate your acceptance of this License to do so, and
all its terms and conditions for copying, distributing or modifying
the Program or works based on it.

<LI>

Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the
original licensor to copy, distribute or modify the Program subject to
these terms and conditions.  You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties to
this License.

<LI>

If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent issues),
conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License.  If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Program at all.  For example, if a patent
license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
refrain entirely from distribution of the Program.

If any portion of this section is held invalid or unenforceable under
any particular circumstance, the balance of the section is intended to
apply and the section as a whole is intended to apply in other
circumstances.

It is not the purpose of this section to induce you to infringe any
patents or other property right claims or to contest validity of any
such claims; this section has the sole purpose of protecting the
integrity of the free software distribution system, which is
implemented by public license practices.  Many people have made
generous contributions to the wide range of software distributed
through that system in reliance on consistent application of that
system; it is up to the author/donor to decide if he or she is willing
to distribute software through any other system and a licensee cannot
impose that choice.

This section is intended to make thoroughly clear what is believed to
be a consequence of the rest of this License.

<LI>

If the distribution and/or use of the Program is restricted in
certain countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Program under this License
may add an explicit geographical distribution limitation excluding
those countries, so that distribution is permitted only in or among
countries not thus excluded.  In such case, this License incorporates
the limitation as if written in the body of this License.

<LI>

The Free Software Foundation may publish revised and/or new versions
of the General Public License from time to time.  Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.

Each version is given a distinguishing version number.  If the Program
specifies a version number of this License which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation.  If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software
Foundation.

<LI>

If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the author
to ask for permission.  For software which is copyrighted by the Free
Software Foundation, write to the Free Software Foundation; we sometimes
make exceptions for this.  Our decision will be guided by the two goals
of preserving the free status of all derivatives of our free software and
of promoting the sharing and reuse of software generally.



<P><STRONG>NO WARRANTY</STRONG>

<LI>

BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
REPAIR OR CORRECTION.

<LI>

IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES.
</OL>


<H2>END OF TERMS AND CONDITIONS</H2>



<H2><A NAME="SEC17" HREF="queue_toc.html#TOC17">How to Apply These Terms to Your New Programs</A></H2>

<P>
  If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.


<P>
  To do so, attach the following notices to the program.  It is safest
to attach them to the start of each source file to most effectively
convey the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.



<PRE>
<VAR>one line to give the program's name and an idea of what it does.</VAR>
Copyright (C) 19<VAR>yy</VAR>  <VAR>name of author</VAR>

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
</PRE>

<P>
Also add information on how to contact you by electronic and paper mail.


<P>
If the program is interactive, make it output a short notice like this
when it starts in an interactive mode:



<PRE>
Gnomovision version 69, Copyright (C) 19<VAR>yy</VAR> <VAR>name of author</VAR>
Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
type `show w'.  This is free software, and you are welcome
to redistribute it under certain conditions; type `show c' 
for details.
</PRE>

<P>
The hypothetical commands <SAMP>`show w'</SAMP> and <SAMP>`show c'</SAMP> should show
the appropriate parts of the General Public License.  Of course, the
commands you use may be called something other than <SAMP>`show w'</SAMP> and
<SAMP>`show c'</SAMP>; they could even be mouse-clicks or menu items--whatever
suits your program.


<P>
You should also get your employer (if you work as a programmer) or your
school, if any, to sign a "copyright disclaimer" for the program, if
necessary.  Here is a sample; alter the names:



<PRE>
Yoyodyne, Inc., hereby disclaims all copyright
interest in the program `Gnomovision'
(which makes passes at compilers) written 
by James Hacker.

<VAR>signature of Ty Coon</VAR>, 1 April 1989
Ty Coon, President of Vice
</PRE>

<P>
This General Public License does not permit incorporating your program into
proprietary programs.  If your program is a subroutine library, you may
consider it more useful to permit linking proprietary applications with the
library.  If this is what you want to do, use the GNU Library General
Public License instead of this License.


<P>
<A NAME="IDX135"></A>


<H1><A NAME="SEC18" HREF="queue_toc.html#TOC18">Index</A></H1>

<P>
Jump to:
<A HREF="#cindex_+">+</A>
-
<A HREF="#cindex_-">-</A>
-
<A HREF="#cindex_[">[</A>
-
<A HREF="#cindex_a">a</A>
-
<A HREF="#cindex_b">b</A>
-
<A HREF="#cindex_c">c</A>
-
<A HREF="#cindex_d">d</A>
-
<A HREF="#cindex_e">e</A>
-
<A HREF="#cindex_f">f</A>
-
<A HREF="#cindex_g">g</A>
-
<A HREF="#cindex_h">h</A>
-
<A HREF="#cindex_i">i</A>
-
<A HREF="#cindex_j">j</A>
-
<A HREF="#cindex_k">k</A>
-
<A HREF="#cindex_l">l</A>
-
<A HREF="#cindex_m">m</A>
-
<A HREF="#cindex_n">n</A>
-
<A HREF="#cindex_o">o</A>
-
<A HREF="#cindex_p">p</A>
-
<A HREF="#cindex_q">q</A>
-
<A HREF="#cindex_r">r</A>
-
<A HREF="#cindex_s">s</A>
-
<A HREF="#cindex_t">t</A>
-
<A HREF="#cindex_v">v</A>
-
<A HREF="#cindex_x">x</A>
<P>
<H2><A NAME="cindex_+">+</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX106">+ as wildcard, qsh</A>
</DIR>
<H2><A NAME="cindex_-">-</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX53">--half-pty|--full-pty|--no-pty</A>
<LI><A HREF="queue.html#IDX58">--help</A>
<LI><A HREF="queue.html#IDX45">--host hostname</A>
<LI><A HREF="queue.html#IDX49">--immediate|--queue</A>
<LI><A HREF="queue.html#IDX47">--robust-host hostname</A>
<LI><A HREF="queue.html#IDX57">--version</A>
<LI><A HREF="queue.html#IDX55">--wait|--batch</A>
<LI><A HREF="queue.html#IDX116">-DHAVE_IDENTD compile-time security option, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX44">-h hostname</A>
<LI><A HREF="queue.html#IDX46">-H hostname</A>
<LI><A HREF="queue.html#IDX48">-i|-q</A>
<LI><A HREF="queue.html#IDX52">-o|-p|-n</A>
<LI><A HREF="queue.html#IDX56">-v</A>
<LI><A HREF="queue.html#IDX54">-w|-r</A>
</DIR>
<H2><A NAME="cindex_[">[</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX51">[--spooldir spooldir]</A>
<LI><A HREF="queue.html#IDX50">[-d spooldir]</A>
</DIR>
<H2><A NAME="cindex_a">a</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX37">Access control list, host</A>
<LI><A HREF="queue.html#IDX81">apparant load average</A>
<LI><A HREF="queue.html#IDX133">author, primary</A>
</DIR>
<H2><A NAME="cindex_b">b</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX65">batch queue, configuring</A>
<LI><A HREF="queue.html#IDX126">Bug List</A>
<LI><A HREF="queue.html#IDX125">Bug reports</A>
</DIR>
<H2><A NAME="cindex_c">c</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX32">Cluster-wide usage, setting up</A>
<LI><A HREF="queue.html#IDX40">command line options, <CODE>queue</CODE></A>
<LI><A HREF="queue.html#IDX24">configure</A>, <A HREF="queue.html#IDX30">configure</A>
<LI><A HREF="queue.html#IDX25">configure, options, installation by ordinary user</A>, <A HREF="queue.html#IDX31">configure, options, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX68">CPU power differences, defining</A>
</DIR>
<H2><A NAME="cindex_d">d</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX128">Development List</A>
<LI><A HREF="queue.html#IDX131">Development projects</A>
<LI><A HREF="queue.html#IDX8">DISPLAY</A>
</DIR>
<H2><A NAME="cindex_e">e</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX5">EMACS</A>
<LI><A HREF="queue.html#IDX76">exec</A>, <A HREF="queue.html#IDX86">exec</A>
</DIR>
<H2><A NAME="cindex_f">f</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX124">Feedback</A>
<LI><A HREF="queue.html#IDX83">formula, apparant load average</A>
</DIR>
<H2><A NAME="cindex_g">g</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX127">Getting Help</A>
</DIR>
<H2><A NAME="cindex_h">h</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX115">HAVE_IDENTD compile-time security option, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX23">home directory, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX129">Homepage</A>
<LI><A HREF="queue.html#IDX79">host</A>
<LI><A HREF="queue.html#IDX38">Host access control list</A>
</DIR>
<H2><A NAME="cindex_i">i</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX118"><CODE>ident.c</CODE>, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX122">identd</A>
<LI><A HREF="queue.html#IDX117"><CODE>identd</CODE>, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX135">Index</A>
<LI><A HREF="queue.html#IDX9">install, ordinary user</A>
<LI><A HREF="queue.html#IDX28">installation as root</A>
<LI><A HREF="queue.html#IDX21">installation by ordinary user, home directory</A>
<LI><A HREF="queue.html#IDX20">installation by ordinary user, NFS directory permissions</A>
<LI><A HREF="queue.html#IDX27">installation by system administrator</A>
<LI><A HREF="queue.html#IDX29">installation by system administrator, NFS directory setup</A>
<LI><A HREF="queue.html#IDX26">installation, preferred method</A>
<LI><A HREF="queue.html#IDX14">installation, preferred privileges</A>
<LI><A HREF="queue.html#IDX13">installation, privileges required</A>
<LI><A HREF="queue.html#IDX3">Introduction</A>
</DIR>
<H2><A NAME="cindex_j">j</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX64">job queue, configuring</A>
</DIR>
<H2><A NAME="cindex_k">k</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX132">Krebs, Werner G.</A>
</DIR>
<H2><A NAME="cindex_l">l</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX72">limiting number of jobs</A>
<LI><A HREF="queue.html#IDX73">load average job queue limits</A>
<LI><A HREF="queue.html#IDX82">load average, apparant</A>
<LI><A HREF="queue.html#IDX66">load scheduler, configuring</A>
<LI><A HREF="queue.html#IDX89">loadsched</A>
<LI><A HREF="queue.html#IDX90">loadstop</A>
</DIR>
<H2><A NAME="cindex_m">m</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX77">mail</A>
<LI><A HREF="queue.html#IDX134">maintainer</A>
<LI><A HREF="queue.html#IDX4">Matlab</A>
<LI><A HREF="queue.html#IDX85">maxexec</A>
<LI><A HREF="queue.html#IDX88">maxfree</A>
<LI><A HREF="queue.html#IDX87">minfree</A>
<LI><A HREF="queue.html#IDX6">MIT_MAGIC_COOKIE_1</A>
<LI><A HREF="queue.html#IDX107">MPI host file</A>
<LI><A HREF="queue.html#IDX110">MPI load scheduling algorithm</A>
<LI><A HREF="queue.html#IDX108">MPI Queue job queue</A>
<LI><A HREF="queue.html#IDX100">MPI, running with Queue</A>
<LI><A HREF="queue.html#IDX111">MPI, stage-1 hook</A>
<LI><A HREF="queue.html#IDX101">MPICH</A>
<LI><A HREF="queue.html#IDX16">multiple installations per cluster</A>
</DIR>
<H2><A NAME="cindex_n">n</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX15">network port numbers</A>
<LI><A HREF="queue.html#IDX22">NFS directory permissions, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX93">nice</A>
</DIR>
<H2><A NAME="cindex_o">o</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX39">options, command line, <CODE>queue</CODE></A>
<LI><A HREF="queue.html#IDX2">Overview</A>
</DIR>
<H2><A NAME="cindex_p">p</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX80">pfactor</A>
<LI><A HREF="queue.html#IDX105">plus as wildcard, qsh</A>
<LI><A HREF="queue.html#IDX17">port numbers, network</A>
<LI><A HREF="queue.html#IDX18">port numbers, TCP/IP</A>
<LI><A HREF="queue.html#IDX1">Preface</A>
<LI><A HREF="queue.html#IDX35">profile file</A>
<LI><A HREF="queue.html#IDX63">profile file, configuring</A>
<LI><A HREF="queue.html#IDX112">PVM support</A>
</DIR>
<H2><A NAME="cindex_q">q</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX36">QHOSTSFILE</A>
<LI><A HREF="queue.html#IDX42">qsh</A>, <A HREF="queue.html#IDX102">qsh</A>, <A HREF="queue.html#IDX103">qsh</A>
<LI><A HREF="queue.html#IDX60"><CODE>qsh</CODE> defaults</A>
<LI><A HREF="queue.html#IDX43"><CODE>qsh</CODE>, command line options</A>
<LI><A HREF="queue.html#IDX104">qsh, hostname wild-card character</A>
<LI><A HREF="queue.html#IDX34">queue</A>
<LI><A HREF="queue.html#IDX109">Queue load scheduling algorithm</A>
<LI><A HREF="queue.html#IDX41"><CODE>queue</CODE>, command line options</A>
<LI><A HREF="queue.html#IDX130">queue-tips list</A>
<LI><A HREF="queue.html#IDX33">queued</A>
</DIR>
<H2><A NAME="cindex_r">r</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX69">restrictions, job queue, setting</A>
<LI><A HREF="queue.html#IDX119">RFC 931</A>
<LI><A HREF="queue.html#IDX99">rlimitcore</A>
<LI><A HREF="queue.html#IDX94">rlimitcpu</A>
<LI><A HREF="queue.html#IDX95">rlimitdata</A>
<LI><A HREF="queue.html#IDX97">rlimitfsize</A>
<LI><A HREF="queue.html#IDX98">rlimitrss</A>
<LI><A HREF="queue.html#IDX96">rlimitstack</A>
<LI><A HREF="queue.html#IDX10">root</A>
<LI><A HREF="queue.html#IDX59">rsh</A>
<LI><A HREF="queue.html#IDX61">Running Queue</A>
</DIR>
<H2><A NAME="cindex_s">s</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX62">SAS</A>
<LI><A HREF="queue.html#IDX67">scheduler, configuring</A>
<LI><A HREF="queue.html#IDX120">secure sites, <CODE>identd</CODE>, using Queue with</A>
<LI><A HREF="queue.html#IDX114">security concerns, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX113">security considerations, installation by system administrator</A>
<LI><A HREF="queue.html#IDX70">setting job queue restrictions</A>
<LI><A HREF="queue.html#IDX75">splus</A>
<LI><A HREF="queue.html#IDX11">superuser</A>
<LI><A HREF="queue.html#IDX78">supervisor</A>
<LI><A HREF="queue.html#IDX74">suspending jobs based on load averages</A>
<LI><A HREF="queue.html#IDX12">system administrative privileges</A>
</DIR>
<H2><A NAME="cindex_t">t</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX19">TCP/IP, port numbers</A>
<LI><A HREF="queue.html#IDX121">tcp_wrapper</A>
<LI><A HREF="queue.html#IDX123">tcp_wrapper and ident.c, installation by ordinary user</A>
<LI><A HREF="queue.html#IDX71">times of day job queue restriction</A>
<LI><A HREF="queue.html#IDX91">timesched</A>
<LI><A HREF="queue.html#IDX92">timestop</A>
</DIR>
<H2><A NAME="cindex_v">v</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX84">vmaxexec</A>
</DIR>
<H2><A NAME="cindex_x">x</A></H2>
<DIR>
<LI><A HREF="queue.html#IDX7">xhost</A>
</DIR>


<P><HR><P>
This document was generated on 17 May 2000 using
<A HREF="http://wwwinfo.cern.ch/dis/texi2html/">texi2html</A>&nbsp;1.56k.
</BODY>
</HTML>