UNIX Cluster Usage Policies
Please respect the fact that many people use these systems! We believe that Caltech community members, governed by the Honor System, do not need to be policed. Some information, however, might help you realize what is and is not acceptable to other Caltech community members and what might harm the computer systems that constitute the UNIX Cluster:
- Please do not unplug or turn off any of the UNIX Cluster workstations in the studentlab. Contact the lab attendant if it appears a machine does not appear to be responding. You are encouraged to reboot the computer by logging in with the username reboot if the computer appears to be in a strange state.
- Don't use more space than you need; learn how to use the du command (type man du). Users are currently limited to 60 megabytes. Be sure to read the system messages displayed whenever you log in. The quota may change as the demands on our resources change.
- Don't create and run programs that waste or abuse system resources. Follow the instructions below about running background jobs. Running distributed computation programs is not permitted on the UNIX Cluster. Examples include but are not limited to: the RC5 cryptographic challenge, the SETI@home Project.
In general, be considerate and be honest with yourself. If you think that what you are doing might be unethical or take advantage of other users, either don't do it or contact us at http://help.caltech.edu (request type IMSS-->Other)for an informed opinion. If you are having problems with other users on the system, please try to inform them politely of the problem and try to work things out. If this proves impossible,contact us at http://help.caltech.edu (request type IMSS-->Other).
Long running jobs should be run in the "background" and at a lower priority than regular jobs. To start a job in the background rather than the foreground, simply append an "&" symbol to the end of your command. For example:
% cc silly.c &
If you are running a long job you must nice your job to 19 the highest niceness (and lowest priority) with:
% nice +19 cc silly.c &
Changing the "niceness" of a job
If you forget to set the niceness of your background job or place it in the background after you start it, you should use the renice command to change its niceness. To do this you would need to find out the process id of the job with the ps command then then do:
% renice 19 process-id
Where process-id is the number of process whose priority you wish to change.
Sending stderr to a File
If you run a job in the background and then log out, you should first redirect any error output to a file. Even if your program is written to place all of its output in a file already, it's a good idea. For example:
% nice +19 cc silly.c >& silly.errs &
will compile silly.c nicely and in the background, directing any errors or screen output to a file silly.errs in the current directory. Note: this is valid only for users using the default shell, csh, or csh-compatible shells such as tcsh. If you use another shell you should find out the information about how to nice jobs and redirect the stderr to a file by using the man command to display information about that shell.
Because the cluster is very busy, serving the needs of several thousand users, we must ask that users run only one CPU-intensive job on the Cluster at a time.
Your home directory is on one of several disks attached to the UNIX Cluster file servers along with the home directories of many other users. To find out how much space is left on the disk where your home directory is located, do:
% df /home/username/.
where username is replaced by your login name. The units of this display are in kilobytes.
To find out how much total space you are currently using, do:
% du -s /home/username/.
On SunOS computers the output of this command is reported in kilobytes. On machines running Solaris 2.x you must do du -sk to see your disk usage reported in kilobytes. Only under unusual circumstances should this total be more than a couple of megabytes. The limit is 60MB (60000 kilobytes), but we expect most users to use less than this most of the time.
Also, you can check your account's space usage, by running the UNIX command
while using either telnet or ssh to connect to its.caltech.edu.
A large amount of temporary space has been placed in the directory /ccovol/suntmp. You should create a directory here to contain your files. This space will, for example, be a reasonable place for examining large tar'd archives from FTP sites. Large is not unlimited, so please do not try to keep large amounts of files here for an extended period of time. Files in /ccovol/suntmp not accessed for seven days are subject to deletion as needed to maintain a minimum amount of free space. In other words, seven days in the minimum lifetime of a file, though it may continue to exist beyond that if sufficient room remains available.
Remember, because the scratch space is for temporary storage, you should delete your files as soon as you are done with them. This is particularly important if your file or files are large.
Please note that we will not look kindly on attempts use the /ccovol/suntmp space for quasi-permanent storage through the use of touch or other programs to access your files every time you log in or out. This is explicitly forbidden, and if we discover you doing it, we will disable your ability to log in.