A current, working copy of this document is kept in ~dwinkel/info/login.faq, or its on the web at http://www-personal.umich.edu/~dwinkel/login.html .
The The ITD Login Service and The ITD Statistics and Computation
Service are machines that have been set aside for use by the
university community. To email the administrators
(itd-login-admin@umich.edu).
The Workstations at NUBS Computing Resource Site
are reserved for use from the console. They
are set up so that people can use them for X applications
and other console uses.
The ITD Login Service,
login.itd.umich.edu:
The ITD Statistics and Computation Service, stat.itd.umich.edu:
The ITD Login Service and The Workstations at NUBS Computing
Resource Site are running an operating system called
Solaris 2.5.1. This is also known as SunOS 5.5.1.
The Statistics and Computation Service is also running
Solaris 2.5.1.
All of the machines have been upgraded to Solaris 2.5.1 this
summer.
The alternatives involve getting a joint ITD/CAEN account from the
CAEN Office in the Media Union, which gives you access to the CAEN
Sun Sparc 20s in NUBS and the Media Union Site, and the RS/6000
machines in the Media Union Site.
It is also occassionally possible to get an account on a departmental
unix machine, though not all departments have unix machines, and
not all departments that do have unix machines offer accounts to
all of their students, staff, and faculty.
The login group is soliciting suggestions for packages users
would like installed. If you have a suggestion, either email
it to the login administrators (
itd-login-admin@umich.edu) or me (
dwinkel@umich.edu).
Right now that list includes transcript (including enscript).
It also includes ArcView/ArcInfo.
You can use the ITD Login Service Machines
(login.itd.umich.edu) to run:
-- interactive programs, such as pine, vi, talk;
-- small programs that do not interfere with other applications
running on The ITD Login Service.
The ITD system administrators reserve the right to stop any job
that is "using too much of the computer's resources". Reasons
include:
-- it is taking longer than 10 minutes to produce a result;
-- it is still running after you have logged out.
Use of these machines is restricted to people who have a
subscription to the UMCE Login Service.
In addition, ITD system administrators reserve the right to:
-- remove any files or data left in /tmp or /var/tmp;
-- stop any program running on The ITD Login Service machines that
provides network services to unauthorized users.
You can use the The ITD Statistics and Computation Service
(stat.itd.umich.edu) to run:
-- resource-intensive programs, such as C and Fortran programs
and statistics applications, or any programs that operate on
large datasets.
These machines are set up for handling large amounts of data
and computation. You can also use The ITD Statistics and
Computation Service to
run interactive programs, such as pine, vi, and talk, but the
programs will not be as responsive as they are when they run on
The ITD Login Service.
So that you can more efficiently run programs using large
datasets, ITD provides local disk space on The ITD Statistics
and Computation Service.
You can make use of this space by copying your
datasets temporarily from your IFS home directory to the
local disk space.
Use of these machines is restricted to people who have a
subscription to the UMCE Statistics and Computation Service.
ITD system administrators will remove files that haven't been
modified or accessed in the last 24 hours to prevent local
disk space from filling up and becoming unusable.
ITD system administrators reserve the right to:
-- stop any program running on The ITD Statistics and Computation Service
machines that provides network services to unauthorized
users.
-- stop any program running on the The ITD Statistics and
Computation Service that interferes with applications running
on other network services.
You can use the the NUBS workstations to run:
-- programs, such as SAS, that can take advantage of the NUBS
workstations' graphics capabilities.
-- interactive programs, such as pine, vi, talk;
-- resource-intensive programs that handle large amounts of data
and computation;
Use of these machines is restricted to people who have a
subscription to the UMCE Statistics Service.
ITD system administrators reserve the right to:
-- stop any program, run by a remote user, that interferes with
a program run by a user who is at NUBS.
-- stop any program running on any NUBS workstation that
provides network services to unauthorized users.
Every user on a system running Unix has a unique integer associated with their uniqname. This integer, known as a UID, is used by the system as a shorthand identifier for a user. Most modern versions of Unix can support UIDs with a value from 0 up to about 60,000.
Because we share our password file across UMCE services that access IFS, that means we can have at most 60,000 different uniqnames. Recently, there have been more customers requesting uniqnames than there are available UIDs below 60,000. We haven't been able to create the requested uniqnames until more UIDs become available.
In July, we upgraded to systems that support UIDs greater than 60000, so hopefully, this will not be a problem in the near future. In fact, it shouldn't be a problem on the (now) current system.
There are two main ways to do this on The ITD Login Service. The
easiest is to use the "passwd
The "kpasswd" command will also work to change your password.
You can change your full name (what finger gives out as your full name, and what Pine uses as your full name if you haven't explicitly changed it in Pine) with the command:
chfn
That command will ask you for a new full name, a location (which is also given out by finger), and a phone number. If all you want to change is your full name, then just hit the Return key for the other questions.
There's a fairly simple way to do this, but there is no official
way to do this for a variety of reasons.
The solution is to add these lines to the end of your .login file:
if ( $SHELL == "/bin/csh" ) then
foreach i ($path)
if ( -x $i/tcsh ) then
setenv SHELL $i/tcsh
set shell=$i/tcsh
exec $i/tcsh -l
endif
end
endif
If your configuration has somehow been screwed up, you can get back the standard configuration with the command:
/usr/itd/bin/redot
If you only need a specific configuration file, you can limit it to that file by specifying the file on the command line:
/usr/itd/bin/redot .cshrc
This is usually necessary because someone deleted the files by mistake, or because someone has old files, and they need to be updated. The files affected are: .cshrc, .login, .logout, .xsession, .Xresources, .mwmrc, .principals, .profile, .tvtwmrc, .twmrc, .zephyr.subs, .zephyr.vars, and AppleVolumes.
If you would like finger to return more information, such as a PGP public key, schedule, etc, you should use a .plan file. You can create a .plan file with the commands:
touch ~/Public/.plan
ln -s ~/Public/.plan ~
You can then edit your new .plan file with the command:
pico ~/.plan
Whatever you put in there is what finger will return when you are fingered on the login servers.
For all of the machines, the finger daemon has been replaced with the GNU finger daemon, because of performance issues with the Solaris finger daemon, and the number of users that we have. The output of this finger program is a little different, as is how the program works. Use "finger -l" for long output that includes the .plan file.
We are currently using a modified version of Gnu's fingerd on The ITD Login Service to offer finger service to users. Given that, there are some problems with finger in general, and Gnu fingerd specifically:
There are a few methods of doing this.
If you have one of the Internet Connectivity Packages supplied by
the university, you can use Fetch (on the Mac) or Rapid Filer (on
a Windows machine). Connect these to login.itd.umich.edu, using
your uniqname as login, and your password as password.
The second way is to use Kermit. Kermit works much the same as
it did on MTS. When logged into The ITD Login Service, type:
kermit
to get into kermit. If you're transferring ascii or text files,
just use send or receive, and a file name, to send or receive
that file off the login server. Then use Kermit with your local
communications program to finish the transfer. If you're
downloading a file from the login server to your local machine
and you use Procomm Plus, it might look something like this:
stargate% kermit
C-Kermit> send fyle
The best method for doing this is to change it in your .login file. Just add the following line to the end of your .login file:
setenv EDITOR vi
That will change the default editor to vi for you.
This is generally not recommended, because the message of the day is there to inform you of changes, and important bits of info on The ITD Login Service. However, if you really don't want to see this message, there is a fairly simple way to make it go away. The way you disable the message of the day is with a file in your home directory called ".hushlogin". You can create this file with the command:
touch ~/.hushlogin
You will then no longer see the message of the day. If you want to reenable the message, just remove the file with the command:
rm ~/.hushlogin
I was curious about this question, because of the number of people who like printing from pine on The ITD Login Service to a local printer, so I wrote a little script that prints a file to the local printer. Currently, to run it, use the command:
~dwinkel/bin/ansiprint
On a Mac, this will bring up the printer dialogue, and ask you if you're sure. On a PC, it may do any of several things. If it just scrolls the message on your screen, then your communications software doesn't support this feature, and you'll have to download the file and print it normally.
The ITD Login Service, The ITD Statistics and Computation Service,
and The Workstations at NUBS Computing Resource Site
can also print to the following printer names:
angell Angell campus computing site HP 4Si queue
angell4q Angell campus computing site HP 4Si queue
angell3q Angell campus computing site HP 3Si queue
dana Dana campus computing site
frieze Frieze campus computing site
sph School of Public Health campus computing site
nib School of Nursing
nubs NUBS campus computing site
largejobs NUBS campus computing site large job queue
seb SEB campus computing site
ugl UGLI campus computing site
union Union campus computing site
To print a file to a nubs printer, for example, use the command:
The resize command in Solaris 2 isn't quite as useful as the one in the previous version of the OS, so what I would do is create a script in a bin directory... And use that. To do that, I would do the following:
Make a bin directory if you don't have one:
mkdir ~/bin
Edit a file called "~/bin/rs", and add the following lines:
#!/bin/csh
eval `resize`
Then change the permissions on the file to make it executable
with the command:
chmod 700 ~/bin/rs
Then, every time you find yourself with the wrong screen size,
just type "rs", and it will resize the screen for you.
Note that the quotes in the above command are back-quotes, not
normal apostrophes.
You can also accomplish the same thing by adding this line to the end of your ~/.cshrc file:
alias rs eval `resize`
Check
http://www.umich.edu/how-to-homepage.html. For more info
on www, check the umich.www newsgroup, and the FAQ for that
newsgroup in:
http://www.us.itd.umich.edu/umich-www/umich-www-faq.html
If the person is at the University of Michigan, you can look up their email address in X.500 with the "ud" command. When you type "ud", you get a * prompt. Some basic commands at the * prompt are:
display
--- displays the information for that person.
display
--- displays the info for that person, or if that matches
more than one person, it will give you a list of people
to choose from.
modify
--- This is how you change information in your X.500 entry.
There are a lot more things that you can do in "ud". If you need
to do more, your best bet is to glance through
one of the pieces of X.500 documentation in the sites.
finger
If the person you're looking for is NOT at the University of Michigan, then the next place to check is in gopher. if you type "gopher" at the % prompt, it will come up with a list of choices. Choose Phone Books, and go on from there... If you can't find the person's address there, you'll have to call them up and ask them directly.
The way to do this is with a (fairly) simple command:
fs lq ~
That will give you something like this:
stargate% fs lq ~
Volume Name Quota Used % Used Partition
user.dwinkel 30000 22559 75% 89%
This is now possible to accomplish over the computer. The
quickest way to do this is to (on the login machines) type:
umce
From there, choose "UMCE Subscription Service", down to "Change
IFS Subscription", and choose that.
Then change the "Quantity to Subscribe", and choose the
"Perform Change" text. Then you'll have more (or less) IFS space.
You can also reach this from the umce.itd.umich.edu host.
When you telnet from the "Host:" prompt on the NAS, you are asked for authorization. Your uniqname will work as authorization, with a twist. You have to include the "domain". So, for me, that would be:
dwinkel@umich.edu
(And then my uniqname password as password...)
This is done because the NAS dialin lines are administered by a
group that has to make sure only authorized customers use the
modems, and there isn't a way for the login servers to get that
information back to them. So, they require their own login
sequence.
It looks something like this:
Host: login.itd.umich.edu
Access Controlled
login: dwinkel@umich.edu
Password? *****
Connected to stargate.rs.itd.umich.edu
login: dwinkel
Password? *****
stargate%
You can use The ITD Login Service to download from the archives. If you ftp to login.itd.umich.edu and change directory to the "~archive" directory, you are looking at the archives. You can also download site licensed software this way. That software is in the "~swdist" directory.
If you would like to use Gnu utilities for compiling, and for general environment feel, you can do it. Just add the following line to your .cshrc file:
set path=( /usr/um/gnu/bin $path )
These permissions are the afs Access Control Lists, or ACLs. You can set (and list) these with the "fs" command. To list the ACL for a particular directory, use the command:
fs listacl
To set the ACL for a directory, use:
fs setacl
where "
Pine and mh are the two email packages on The ITD Login Service.
This problem USUALLY means that your IFS Home Directory is over its quota. You can check this with the command:
fs lq ~
If you're "Used" portion is more than your "Quota" portion, then you are over your quota. You'll either have to make your quota larger ('lynx http://www.ifs.umich.edu/~ifs'), or delete some files to fix the problem.
If you want to do this, all you have to do is create a file
called .signature in your home directory, edit that file, and
save it. That file will get added to the end of every message
you send using pine, if you don't delete it.
To actually create the file, you can use the pico
editor. You can do this with the command:
pico ~/.signature
Then put the text you want into that file, and quit the editor by hitting the Control key and X. Then say you want to save the file, and hit return to accept the default name. If you want to change it later, you can do that with the same command!
To do this, you have to change some items in your pine config. Go to the Main menu in Pine. From that, choose "Setup" by pressing the "S" key. Then choose "Config" by pressing the "C" key. Then you'll notice a section called "folder-collections". That section we want to change so that it looks like this:
folder-collections = mail/[]
IMAP {"u".imap.itd.umich.edu:143}mail/[]
(Where "u" is the first letter of YOUR uniqname)
The next time you use pine, it will save (by default) to the
first folder collection listed. If that's not the collection
you want, then change it back in configuration.
MH is available on The ITD Login Service in /usr/um/bin. Currently, it is unsupported, meaning that ITD makes no guarantee that it will continue to be available, consultants can't answer questions about it, and there is no ITD-specific documentation for it. ITD system administrators are working to make it a supported application so that customers will have reasonable documentation and consultants can answer customer questions.
MH on the login servers is configured to use KPOP to get mail from the IMAP machines. You can tell MH where your maildrop is by setting the MAILHOST environmental variable to be the domain name of your IMAP host. For instance, user cel might type:
setenv MAILHOST c.imap.itd.umich.edu
^ ('c' is the first letter of the uniqname 'cel')This is a command you might want to add to your .cshrc or .profile. After typing this command, you can use the "inc" command to receive mail from IMAP mailbox into your +inbox folder. "man mh" will display a summary of MH commands that you can use to begin exploring MH. O'Reilly publishes an excellent nutshell handbook on MH, available at Borders and Barnes and Noble bookstores.
Be careful about switching between a POP mail client like MH and an IMAP mail client like pine. POP clients can only see mail items in IMAP's INBOX folder, and can only store mail in your IFS directory. If you want to move mail from other folders in your IMAP mailbox, forward the mail items to yourself, and inc the mail. Likewise, to get mail from IFS MH folders back to your IMAP mailbox, forward the mail to yourself, and process it with an IMAP client like pine.
Mail, UNIX Login, and many other services are being provided on
a subscription basis. You pay for the service at the beginning
of the month, and barring running out of disk space, you can use
it as much as you want. From the first message to the 10000th,
it's the same price.
When we were charging on a per-unit basis (i.e. CPU accounting,
etc), it made sense to separate your mail and other activities
based on what you are doing. In the current environment, this
is less of an issue.
Departmental allocations are still present, although they are
smaller than in the past, due to the increased allocations to
individuals. Probably many departments will spend their
allocations on things like LAN services, more expensive email
options, and other, non-basic services.
The term "account" here is a little misleading. Your account
in the UMCE is the place where you have money to pay for
services, such as ConferU, IMAP, IFS, UNIX Login, etc. You may
have several accounts (i.e. several funding sources, such as
personal [self-funded], request [monthly allocation], and
departmental), but you have one uniqname, which is associated
with, most likely, one email mailbox. So all of your mail goes
to one place.
Basically, the answer is "use mh" for the ITD provided
email alternatives, though if your department provides email
access, one or both of the following may work for you:
There are tools available on most UNIX systems, for
example: procmail and MH. I will address both of them here.
Procmail is designed to be used in the model where you log into
the same machine that contains your mail, either on its local
disk, or its mail partition imported in some fashion. You set
up a .forward file that pipes your mail through procmail, which
looks at a configuration file to find out what to do with your
various messages.
This does NOT work on the login.itd.umich.edu machines, and
never will, both because procmail is a finicky program it's
easy to lose email with, and because The ITD Login Service isn't
related to The IMAP Email Service, other than providing a method
(pine or mh) for accessing it.
MH, the Rand Mail Handler, doesn't process your mail when it
comes it, but when you read it. This is available on The ITD
Login Service.
As for other mail handlers, they aren't being used for various
reasons. Most of them aren't being used because the protocols
they use to handle mail don't scale well to a user base of tens
of thousands. Both Berkeley mail and elm suffer from these scaling
problems.
Again, because the mail isn't stored on a "local" file system to the login servers, this doesn't happen automatically. If you would like to have it happen for you, though, you can add this line to your ".login" file:
imapcheck -v
This will not only tell you if you have new mail, but also how many new messages you have waiting for you when you run pine.
For The Workstations at NUBS Computing Resource Site, there is also a program called "xmailbox" that should put a mailbox with the number of new messages in your mailbox below it, and will check your mailbox for new mail on a regular basis. To run this, just type:
xmailbox &
at a unix prompt, while at the console of one of the NUBS work- stations.
Get PGP, compile PGP, and run PGP...
Otherwise, if you're not quite that paranoid, PGP-2.6.2 is on the
login servers in /usr/um/bin and /usr/um/pgp-2.6.2. Check it out.
Yes, there are several other secure features to the login servers. The login servers run kerberized telnetd, and kerberized telnet, which means that, with a Kerberized telnet client (for example, nifty-telnet on the Mac, or ktelnet on a unix machine), you can telnet to a login machine with a completely encrypted link. To get an encrypted session to a login machine without having Kerberos installed, you can use ssh. ssh also works to telnet between the login servers, and between login servers and the confer.itd.umich.edu machine. What this means is that you can type the following command:
ssh confer
to get a prompt on the confer machine without retyping your uniqname or password. The login team is also experimenting with kerberized ftp, though I'm not sure of any desktop clients that support kerberized ftp.
This is generally discouraged, because newsreading in pine is buggy, and occasionally messes up newsgroups. tin is suggested instead, but if your heart is set on pine:
Go to the Main menu, type 'S' for Setup, and 'C' for Config.
Then use the arrow keys to get down to "nntp server". Hit 'A'
to add a value, and specify "news.itd.umich.edu" as your server.
Then hit
As with many things in the new environment, this is up to your own personal preference. Most people prefer to read news with tin, though. This is because tin offers a fairly easy interface (that's a lot like pine's) for reading news in an organized way. trn, rn, and nn are also available for those who like those news readers.
As perhaps one of the most common questions about uselessnet, I
can't (at this moment) think of a good answer to this question,
so I'm going to let it stew in my mind for a while.
Of course, you can also look at one of the FAQ's from the binaries
newsgroups. They should get posted at least once a month.
The DISPLAY variable is the variable every X program checks to see where the information it is going to display should be sent. So, for example, if I'm logged into biscuit.rs.itd.umich.edu at the NUBS Computing Resource Site, and I'm running SPSS on stat.itd.umich.edu, then I should have the DISPLAY variable in the session I'm running SPSS from (on stat.itd.umich.edu) set to biscuit:0. I can do that with the command:
setenv DISPLAY biscuit:0
This will tell the stat.itd.umich.edu machine that I want the SPSS
windows displayed on biscuit, using display number 0, which is the
only display biscuit has.
This is not all that is needed, however. I also need to set it up
so that biscuit will accept display data from stat.itd.umich.edu.
To do this, I need to type more commands. If I open a new shell on
biscuit, and type this command:
xauth extract .Xauth biscuit:0
And then go back to my shell on stat.itd.umich.edu, and type the following command:
xauth merge .Xauth
Then it will work. There is a method of X "security" using the xhost command, but that method is easily subject to snooping, so it is not recommended.
First, you will have to get an X-Windows server set up on your
PC. Many people are using Xceed for Windows, or MacX for the
Macintosh to do this.
Next, you will have to set your DISPLAY variable as above, to your
desktop PC. The command should look something like this:
setenv DISPLAY your.machine.here.umich.edu:0
This command should have your machine's name, or IP number in it
before the colon (:). If you don't know what your machine name is,
you can find out (when logged into a login or statistics server) with
the command who | grep <uniqname>. When you
type that command with your uniqname instead of "<uniqname>",
your machine's name will be in paretheses. If
all that's there is a series of numbers, don't worry, they will
work as well. Then you will have to figure out how to allow
the machine you are logged into to open windows on your desktop
machine. I, unfortunately, don't have directions for doing this with
Xceed or MacX, though it should be almost as easy as dragging down a
menu, and entering the hostname of the host you're logged into.
Last, just run an X application. It should work just fine for you,
if everything is set up right. If everything isn't set up right,
there are a few places you can go for help. You can call 764-HELP,
or send a message to online.consulting@umich.edu, talk to your
departmental network admin, or you can contact
itd-login-admin@umich.edu. This type of use is slow over dial-up
links, but can be acceptable for some things, especially over an
ethernet connection.
As with everything else, there are security concerns with X. The first of those is that data sent over an X connection isn't encrypted in any special way, so typing your password into an X window that's running on a different machine can allow for someone to break your password. Using xauth minimizes this danger. If you choose to use xhost or use X-Windows from your PC, this is more dangerous, because they don't do the encryption and decryption that the xauth method uses. As with everything, changing your password fairly often will minimize your risk.
There are a few reasons to run SPSS via X-Windows on the Statistics servers. The primary reasons have to do with the local disk that's on the Statistics servers, and the speed with which the Statistics servers can crunch data. What using X-Windows to display SPSS windows on a desktop PC from a Statistics server does is, it allows you to use the power of the statistics server from your desktop.
set path = ( ~/bin/$CPU ~/bin /usr/local/bin /usr/local/etc \
/usr/vice/bin /usr/um/bin /usr/itd/bin /opt/SUNWspro/bin \
/usr/ccs/bin /usr/bin /usr/ucb /usr/bin/X11 . )
There are three "versions" of C on The ITD Login Service.
- gdb can't debug binaries created from the standard cc
If you would like to use Gnu utilities for compiling, and for general environment feel, you can do it. Start by adding the following line to the end of your .cshrc file:
set path=( /usr/um/gnu/bin $path )
Then, add the following lines to the end of your .login file:
set path=( /usr/um/gnu/bin $path )
setenv MANPATH /usr/um/gnu/man:$MANPATH
This should set you up to use all of the gnu utilities rather than the Sun-supplied complements. Please be aware that /usr/um/gnu/bin also contains replacements for commonly used Unix utilities such as find, ls, and grep.
To do this, I believe someone would have to write a modified version of the file /usr/ucb/cc which instead invokes gcc. Naming it, and finding a place for it would be two additional steps needed. It may make sense for diehard BSD fans to put such a modified script in their own ~/bin directory. That plus setting your paths as in section 5.1.4 should be about enough.
There are two different C++ compilers currently available.
- The compiler CC (/opt/SUNWspro/bin/cc) is the Sunsoft version
4.0 C++ compiler, and is a system V compiler environment.
It includes headers from /usr/include, and uses libraries out
of /usr/lib. The header files in /usr/include are written to
be compatible with both Sunsoft C++ and C.
To get man information, type man CC.
This is a currently Sun supported product, which we have a
support contract for. It will evolve moderately, but should be
available in this basic form for at least the next 5-10 years.
- another C++ compiler is /usr/um/bin/g++, which is
the GNU C++ compiler. This compiler, unlike the native System
V C++ compiler, searches the GNU C++ header files first, which
are stored in /usr/um/gcc-2.7.2/lib/g++-include
To get man information, type man g++
To get GNU info pages about GCC, type info -f /usr/um/gcc-2.7.2/info/gcc
To get GNU info pages about libg++, type:
info -f /usr/um/gcc-2.7.2/info/libg++
To get information about the GNU info command, type man info
The program info is a GNU hypertext documentation viewer that
has an interface resembling emacs.
The GNU g++ compiler should be installed and operational at
this time (Oct 2, 1995), so feel free to use it, and report
problems the normal way.