EECS Instructional Support, University of California at Berkeley
[ ISG home page ] [ who we are ] [ send us email ] [ search ] [ FAQ ]

                        College of Engineering
            EECS Instructional & Electronics Support Groups


/share/b/pub/reports/managers/Fall_2000


								Jan 10, 2001

              EECS Instructional Computing - Review and Plans
	      -----------------------------------------------
		                  Fall 2000


References about Common ISG and ESG Services:
--------------------------------------------

  http://inst.eecs.berkeley.edu/
    Instructional WEB server, links to class home pages, student home 
    pages, information about Instructional UNIX accounts, modem access, 
    cardkey access, computers and labs, software.

  http://iesg.eecs.berkeley.edu/
    Electronics Support WEB server, links to information about 
    electronics labs, AV services and Windows NT services.


Recent improvements to Instructional labs:
------------------------------------------

  1) SunRay 'xterminals'  (310 Davis, 271 & 273 Soda)

     45 new SunRay 'xterminals' were installed in 310 Davis for CS61A.
     (See http://www.sun.com/products/sunray1/).  The lab is served by 
     a new Sun SPARC E450 server running Solaris.  This equipment was 
     purchased at a discount from Sun and replaces aging HP workstations.

     66 SunRays were donated by Sun in January 2000 and were installed in
     271 and 273 Soda.  These labs are served by 2 E5000 servers that were
     donated by Prof Culler's research group.

  2) Intel and Microsoft Grants received in July 2000

     20 new Intel PC running Windows 2000 were installed in 123, 218 and 
     264B Cory for use by EECS150, EE1, EE143 and other classes.  

     30 new Intel PCs running Windows 2000 were installed in 330 Soda for use 
     by CS162, CS169, CS172, CS186 and other classes for drop-in use.

     Intel also donated a new Dell Quad-PIII Xeon Server that is being 
     installed to support new Instructional IMAP and WEB services.

     Microsoft donated a system to run the Instructional Exchange server.

  3) More NT 'drop-in' seats in Cory Hall

     24 NT systems are now available in 111, 117 and 199 Cory for drop-in 
     use by students in EE and CS classes.   CS160 will sometimes reserve 
     117 Cory.  An additional 15 NT systems in 119 Cory are sometimes 
     available for drop-in, although EE290T and CS152 students will reserve 
     the lab at times.

  4) Intel "Nomadic" Lab (117 Cory)
   
     8 IBM ThinkPad docking stations are now installed in 117 Cory, for use
     by CS160 students.  With a grant from Intel, Profs Joseph and Demmel 
     obtained 8 new monitors/keyboards/mice as well.   CS160 students have
     been issued the laptops and wireless ethernet cards, and they are
     using the nacent EECS wireless network.

  5) EECS CD for Students

    Instructional staff are working with several CS faculty to assemble a 
    collection of free software that is commonly used by students for use 
    on their home computers.  Instruction will initially spend about $1000 
    to have 1000 CDs burned, for free distribution to our students.   This 
    is done now with floppy disks that are given from 387 Soda.  Perhaps 
    we'll collect something we need in trade for the CDs (maybe a ream of 
    printer paper!).

    The CD will include things that are not on IS&Ts Connecting@Berkeley CD 
    but are already freely available: TereTerm SSH and some variants, scheme, 
    spim, emacs, MacX, GNU and CynWin compilers and maybe java.  The idea is 
    to encourage use of home computers to help offload the increasing demand 
    for our labs.

 6) New printers

    3 new HP 8000DN printers (24 duplex, b&w pages/minute) have been installed 
    in 274 Soda for shared use from the Instructional labs.   An older printer 
    was moved to 273 Soda.  These have a combined bin capacity of 10,000 sheets 
    of stored paper and a combined printing capacity of 96 pages per minute.

    A second HP 8000DN printer is being added in 199 Cory as well, for combined 
    bin and printing capacities of 2,000 sheets and 48 pages/minute.


Current Issues Related to Instructional Computing:
-------------------------------------------------
  
  1) Inceasing enrollments vs decreasing lab space:

     To accommodate increasing enrollments in all UC campuses the Office
     of the President have announced a program called "Tidal Wave II". 
     UCB enrollment will increase by 4,000 students (15%) in the next 
     decade.  (references: http://www.ucop.edu/planning/lrenroll.html and 
     http://www.ucop.edu/planning/enrollpres0100/sld007.htm).

     At the same time, there are plans to convert Instructional lab space
     in Soda hall for use by new faculty research.   30 seats (330 Soda)
     will be vacated in Fall 2001, and as many as 130 seats (271, 273, 275,
     277 Soda) will be vacated later, when equivalent space will become
     available in the refurbished Hearst Mining building.  (Reference: 
     http://www.cs.berkeley.edu/soda.minutes.shtml).

     In addition, CS3 will vacate 175 LeConte in May 2001 and will need a
     new lab of at least 30 seats.  We are porting their software (Gambit)
     from MacOS to UNIX.  This is still being planned, but we anticipate 
     that CS3 will be relocated to 1107 Etcheverry or 349 Davis.

     Current lab allocation is summarized in 
	http://inst.eecs.berkeley.edu/~iesg/iesglabs.html

  2) Enabling Students to Work from Home

     One way to ease the space crunch is to make it easier for students to
     work from home and via mobile laptops.   Increasingly, dorm rooms and 
     apartments are wired with high speed networks.

     Some initiatives:

- downloadable software (http://inst.eecs.berkeley.edu/software)
     - provide "ssh" for free, for MS Windows, UNIX and MacOS
     - obtained donation of Visual C++ and J++ for students to take home
     - new Instructional software CD 
     - IS&T "Connecting@Berkeley" CD
     - network ports for laptops will be installed in some labs

  3) Improvements Needed in 310 Davis 
    
     310 Davis has been reserved for CS61A for the last 3 years.  Instruction
     has invested $20,000 in new network and security features and $50,000 for
     new computers (see SunRays, above).  
     
     But 310 Davis lacks 24-hour bathroom access, is run-down and is often 
     overheated.  Bathroom access into Davis Hall will be possible when a new 
     cardkey reader is enabled (in October?).  PD&C has quoted $39K to install 
     A/C.

     Additional improvements needed: carpet, paint, drop ceiling.


Notable events this semester: 
-----------------------------

  ----------------------------------------------------------------------------
  Nov 27 - SAIDAR will be rebooted from 11:45am - noon to clear a 
	problem that is preventing access to users' email and files from 
	some computers.
  ----------------------------------------------------------------------------
  Oct 18 - Saidar.eecs is being rebooted, 9:50-10:10am today

	We had to reboot Saidar to clear a problem that was preventing
	access to the secure WEB server (https://saidar.eecs.berkeley.edu).

	We are working to move home directories off of Saidar so that these 
	problems will not affect the users.   The problem seems to be an 
	incompatibility between the older NFS automounter daemon on Saidar 
	and the newer NFS deamons on other servers (Mamba, the SWW) from which
	Saidar mounts some filesystems.   We are also changing Saidar for now
	so that it does not use the automounter.
  ----------------------------------------------------------------------------
  Oct 12 - Moving home directories from saidar.eecs to mamba.cs

	We have started to move UNIX home directories from the file server
	saidar.eecs to mamba.cs.   We have added disk space to mamba.cs and
	will centralize the Instructional UNIX home directories there over
	time.    Users cannot login directly to either fileserver.

	You can identify your home directory server and filesystem with the
	UNIX command "df -k ." These are the Saidar filesystems:

		/home/b, /home/c, /home/d, /home/e, /home/f

	The new filesystem on Mamba is called /home/aa.

	We moved 766 accounts on Tuesday Oct 10.   These were accounts that
	had been inactive for several weeks, so we sent no warnings.

	These are the only effects that users might see:

	- Loss of access to the home directory for about 30 minutes while 
	  it is being moved.   We will send warnings several days early by 
	  email from now on.

	- Any use of a hard-coded directory path to files in your home 
  	  directory will have to be changed.   That is, if you have the old
	  path "/home/b/users/mylogin" in your .cshrc, you'll need to change
	  it to "/home/aa/users/mylogin".  This is most likely to occur in 
	  your .cshrc and any shell scripts or CGI scripts that you may have 
	  written.    We will search for such cases and warn the user if we 
	  find them.
  ----------------------------------------------------------------------------
  Oct 12 - Saidar.eecs was down, 9:15am-9:35am today
  Oct 10 - Saidar.eecs was down, 1:38pm-2:00pm today

	We had to reboot Saidar to clear a problem that was preventing
	access to the secure WEB server (https://saidar.eecs.berkeley.edu).
  ----------------------------------------------------------------------------
  Oct 06 - Mingus.eecs (WebCAD, SUPREM, etc.) was down, 1:00 - 1:30 pm today
  ----------------------------------------------------------------------------
  Oct 03 - Saidar.eecs was down, 9:45am-10:45am today

	The major symptoms were login sessions that froze up during startup,
	in the Instructional labs (310 Davis, 199 Cory, 27x Soda) and via
	'ssh' into the Instructional UNIX servers (cory.eecs, po.eecs, dv310.cs,
	quasar.cs, pulsar.cs, torus.cs).

	The NFS service on saidar.eecs failed, which caused the /share/b and 
	several /home directory filesystems to become unavailable.   We had to 
	reboot Saidar.eecs to clear the problem.
  ----------------------------------------------------------------------------
  Sep 27 - ALL UNIX COMPUTERS DOWN - 6pm-8pm TODAY

	The EECS networking staff have announced that they will replace two
	of the network switches that feed the central departmental servers,
	including the /usr/sww UNIX filesystems and the \\NTSWW NT filesystem.

	When /usr/sww is unavailable, many programs are missing and user logins
	may even freeze up while waiting for one of these programs.  Error
	messages look something like:

	  NFS timeout for server cory: RPC: Timed out

        Where "cory" could be the name of several different servers, such as 
       "cory", "alcyone", "electra", "alcyone".   The server name that appears 
       (such as "cory") may not be the actual server that is unavailable.  The 
       error message may a side effect of some other server being down. 

	So we recommend that users do not use the Instructional UNIX computers 	
	during this time.  These computers include:
		
	labs:		271, 273, 275, 277 Soda
			199 Cory
			310 Davis
	servers:	cory.eecs, po.eecs, mingus.eecs,
			quasar.cs, pulsar.cs, torus.cs, dv310.cs
  ----------------------------------------------------------------------------
  Sep 26 - pulsar.cs is down overnight - disk failure - repaired on Sep 28
  ----------------------------------------------------------------------------
  Sep 25 - Air conditioning failure in Soda Hall labs (Sep 23-25)

	 The Soda Hall Instructional labs, were exceptionally hot and stuffy
	 over the weekend.  A "chiller" failed on Friday, which caused a 
	 radical temperature increase in the building.  It was restarted on
	 Monday morning.

	 There was a failure last summer in the cooling equipment on the roof 
	 of Soda Hall, which has caused a general temperature increase in the 
	 building.  That equipment is scheduled for replacement soon.
  ----------------------------------------------------------------------------
  Sep 15 - 7:30pm-9:45pm - network failure in Soda Hall

	Symtoms on Instructional computers included frozen login sessions
	and "NFS timeout" error messages (loss of access to /usr/sww files).
  ----------------------------------------------------------------------------
  Sep 13 - 5:30pm-6:30pm - network failure between Cory and Soda Halls 

	Symtoms on Instructional computers included frozen login sessions
	and "NFS timeout" error messages
  ----------------------------------------------------------------------------
  Sep 6 - SAIDAR (home directories) was rebooted to clear a problem, 
	down from 12:30pm-12:45pm 
  ----------------------------------------------------------------------------
  Sep 3 - network outage to reload routers

	Network shutdown Sunday @8AM: There will be a quick reboot of routers
	in Soda Hall to switch over to a new UPS. This should not take more
	than 10 minutes.
  ----------------------------------------------------------------------------
  Aug 30 - MAMBA (a ghome directory server) was down from about 6:45pm-7pm
  Aug 29 - SAIDAR (a ghome directory server) was down from about 10am-10:40am

	These were caused by software glitches; we are watching for evidence
	and solutions.

	The effect of this is that home directories from the affected server
	are unavailable, which causes active login sessions to freeze with
	"NFS2 server [computer name] not responding" error messages.   The
	computer name that appears may not be "saidar" or "mamba"; it may be
	the computer you are logged into (such as "cory"), which is failing 
	to reach the file server.
	
	New attempts to login either hang, fail or generate a "using / for 
	home directory" sort of error message.   

	To determine what server has your home directory, login to your EECS
	Instructional UNIX account and type:

		df .		[that's "df", space, period]

	If you see the word "saidar" in the output, then that is the server
	that has your home directory.  ("mamba" is our other home directory
	server for UNIX accounts.)
  ----------------------------------------------------------------------------
  Aug 18 - incoming email and home directories
	 on mamba.cs were down from about 10am Thu - 11am Fri.
  ----------------------------------------------------------------------------
  Aug 18 - incoming email and all home directories
	 will be down from 10pm-10am (Aug 18-19) for full system archiving

	We apologize for the interruption to access to your email and files.

	Overnight tonight (Aug 18-19), we will run full system archives of
	home directories on mamba.cs and saidar.eecs.

	While the home directories are unavailable, we also stop the email 
	server on pasteur.eecs, so that incoming mail is queued and not 
	delivered incorrectly (due to the absence of users' .forward files).  
	Email that is sent to users on our systems (@cory, @pasteur, etc) is 
	cached on another server and will be delivered to you once we restore 
	all the home directories.  

	Also note that the WEB pages for these accounts are not accessible
(URLs such as http://inst.eecs.berkeley.edu/~you)
  ----------------------------------------------------------------------------
  Jul 30 - network problem from about 1am-10am and 4pm-5pm

	There was a network problem from about 1am-10am and 4pm-5pm on Sun 
	July 30 that cut traffic into Soda and Cory Halls.   

	The gigabyte link to the rest of campus and the outside world was reset
	at about 10:05am and the problem was cleared.  EECS networking staff 
	are working with Nortel to identify the cause of the interruption.
  ----------------------------------------------------------------------------
  Jul 26 - home directories on Saidar will be off-line Friday July 28, 8am-10am

	To determine what server has your home directory, login to your EECS
	Instructional UNIX account and type:

		df .		[that's "df", space, period]

	If you see the word "saidar" in the output, then that is the server
	that has your home directory.  ("mamba" is our other home directory
	server for UNIX accounts.)
  ----------------------------------------------------------------------------
  Jun 16 - 10:45-11:15am:  Network equipment failure caused loss of connection 
	to several CS servers

	coeus.cs, the DNS servers and a number of other servers in CS were
	cut off during a network failure

	Symtoms on Instructional computers include frozen login sessions
	and "NFS timeout" error messages
  ----------------------------------------------------------------------------
  Jun 06 - EMAIL delivery to PASTEUR was delayed for 1-2 hours

	while some home directories were off-line for dumping.

  ----------------------------------------------------------------------------
  Jul 1998 - Instructional UNIX password service

	The Password Service
	------------------------------
	The UNIX password software is called "NIS+".  It stores your password
	and other account information in databases on several servers.  To
	change your NIS+ entry, you login (using "ssh") to po.eecs.berkeley.edu
	and enter one of these commands:

	  passwd	- to change your password
	  chsh		- to change your login shell
	  chfn		- to change your name, office address or phone number
	  chkey -p	- to re-sync your "secret key" with your pasword*


	The change propigates to all of the EECS Instructional UNIX computers
	within a few minutes.  For more technical info about NIS+, please type 
	"man nis+"  on any of our Solaris systems (for lists, please see
	http://inst.eecs.berkeley.edu/clients).

      * Your second, hidden password:
	--------------------------------------
	In addition to the 'login' password that you now use, NIS+ uses a 
	second 'secure RPC' (also called 'secret key' and 'NIS+ credential')
	password.  The default 'secure RPC' password for all users is "nisplus".  

	When logging into one of our Solaris X86 computers, you may see a
	warning message such as 

	    This password differs from your secure RPC password.
	or
	    Password does not decrypt secret key for unix.291@Inst.nisplus.

	This is not a problem, but it will be an advantage to you to make your
	'secure RPC' password be the same as your 'login' password.  You can
	do that by typing logging into "po.eecs" or "torus.cs" and running the
	Solaris X86 command "chkey".  For example:

	    % chkey -p
	    Updating nisplus publickey database.
	    Generating new key for 'unix.3232@Inst.nisplus'.
	    Please enter the Secure-RPC password for jdoe: nisplus
	    Please enter the login password for jdoe: {jdoe's password}

        This sets the 'secure RPC' password to match the 'login' password,
	so you won't have to type it when you change your password, shell or
	'finger' information on a Solaris X86 system.  We plan to install 
	"wrapper" programs for the UNIX 'passwd', 'chsh and 'chfn' programs 
	to automate the entry of the 'secure RPC' password, but you may still
	see messages about it when you use these programs.  Please notify 
	'root@cory.eecs' if you have difficulty using these new programs.
  ----------------------------------------------------------------------------


For additional information, please see http://inst.eecs.berkeley.edu or
contact us.

  
  Kevin Mullally, ISG Manager		    Ferenc Kovac, ESG Manager
  EECS Instructional Support Group          EECS Electronics Support Group
  378 Cory Hall, (510) 643-6141             380 Cory Hall, (510) 642-6952
  kevinm@eecs.berkeley.edu                  ferenc@eecs.berkeley.edu
  
  
source: /share/b/pub/reports/managers/Fall_2000