Sun-Spots-Request@RICE.EDU (William LeFebvre) (07/19/88)
SUN-SPOTS DIGEST Monday, 18 July 1988 Volume 6 : Issue 143 Today's Topics: Re: file `which foo` fails on 3.4/3.5 Re: file `which foo` fails on 3.4/3.5 SunOS source licenses (academic) login problem when NFS server crashes (2) Level 2 interrupt problems (3/280) WANTED: Expert System Software/Hardware Info NCAR graphics package for the SUN? Differences between SunOS 3.2 and 4.0? backup software for Suns? UofMaryland television courses: User Interface Strategies '88 Send contributions to: sun-spots@rice.edu Send subscription add/delete requests to: sun-spots-request@rice.edu Bitnet readers can subscribe directly with the CMS command: TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name Recent backissues are available via anonymous FTP from "titan.rice.edu". For volume X, issue Y, "get sun-spots/vXnY". They are also accessible through the archive server: mail the request "send sun-spots vXnY" to "archive-server@rice.edu" or mail the word "help" to the same address for more information. ---------------------------------------------------------------------- Date: Fri, 8 Jul 88 10:06:59 +0100 From: sunuk!sunnl!wizard!dirk@sun.com (Dirk Mansvelder-Sun Nederland TSE) Subject: Re: file `which foo` fails on 3.4/3.5 Mike, It all seems te work over here. Are you sure you did not mix up the quote characters ? file 'which which' gives: which which: No such file or directory while file `which which` gives: /usr/ucb/which: executable c-shell script Regards, Dirk Mansvelder Sun Nederland (e-mail: dirk@NL.sun.com, ..!sun!sunuk!sunnl!dirk) ------------------------------ Date: Fri, 08 Jul 88 04:14:48 -0700 From: Don Kark <wiley!don@uunet.uu.net> Subject: Re: file `which foo` fails on 3.4/3.5 > Anyone know why, on SunOS 3.4 and 3.5, I get > csh% file `which which` > /usr/ucb/which: No such file or directory > csh% file /usr/ucb/which > /usr/ucb/which: executable c-shell script I've seen this problem occur when output is generated in the .cshrc startup file. In my case it had to do with redefining "cd" to be: chdir \!* ; set prompt="$LOG_MACH.$USER ( \! ): "; echo -n "]l$cwd\" Which puts the current directory on the label of the suntools window by echoing the escape sequence to be interpreted by the window <escape seq>. I later did a "cd ." in the .cshrc file. At this point your problem showed up. It turns out that since "which" sources the .cshrc file, any output generated in your .cshrc becomes part of the argument to "file" (even escape sequences). So what "file" sees is an argument "<escape seq>/usr/ucb/which", which is not a valid file. Unfortunately you won't see the complaint correctly because the window interprets the escape sequence! To find the culprit, do an: echo `which which` | od -c It may not always be obvious (perhaps the result of alias/program executions in your .cshrc). I never understood why I never noticed it under 3.2; I always thought I must have introduced the error after going to 3.4. -- Don Don Kark {cit-vax,trwrb,uunet}!wiley!don TRW Inc. wiley!don@csvax.caltech.edu Redondo Beach, CA don@wiley.uucp ------------------------------ Date: 7 Jul 88 21:53:45 GMT From: "P. Ashton" <paul@cantuar.uucp> Subject: SunOS source licenses (academic) We have been in the process of attempting to get a source license (academic-type) for Sun OS for some time now, but things have been progressing very slowly (there's a lot of intermediate layers between us in New Zealand, and people that can answer questions, and make the decisions in the US). What I'd like to know is what flavour of AT&T source license do we need to get a Sun OS source license? We have System III and 32V source licenses from AT&T, which are sufficient for the 4.2 and 4.3 BSD source licences we have (as SunOS was based on 4.2 BSD I figured that if we had sufficient AT&T source licenses to get 4.2 BSD source, then we would have sufficient AT&T source licences to get Sun OS source). We included a copy of our System III license agreement with our application, but I've been unable to ascertain from Sun whether this is sufficient, (we deal with an NZ distributor, who deal with Sun Australia, ....) so I'm asking the net so that if we do need a System V license we can set the wheels in motion to get a System V license. Anyone who has a SunOS source license, or who works for Sun able to shed any light on this? Paul Ashton, from cantuar - the southern most UUCP site. Internet(ish): paul@cantuar.{uucp,nz} JANET/SPEARNET: p.ashton@nz.ac.canty UUCP: ...!{watmath,munnari,mcvax,...!uunet!vuwcomp}!cantuar!paul NZ Telecom: Office: +64 3 667 001 x6350 NZ Post: University of Canterbury, Christchurch, New Zealand ------------------------------ Date: Fri, 8 Jul 88 12:09:13 +0200 From: mcvax!tnosoes!tom@uunet.uu.net (Tom Vijlbrief) Subject: login problem when NFS server crashes (1) Hallo sun-spots readers, We have a 3/160 server which mounts a NFS filesystem from a 3/260 server. Both are running SunOs 3.5, Yellow Pages and the mount is 'soft'. When the 3/260 crashes allmost all is fine, (programs generate 'NFS getattr failed for server sunray: RPC: Timed out' messages), but a login will take about a whole minute, before the shell starts to read .login ! The shell needs nothing from the crashed 3/260 server and does not display any errors... Only root is able to login fast. Killing ypbind does not result in any improvement, so it seems not to be a YP problem. What is happening here? (Umounting the NFS-filesystem fixes the problem but I don't think this would be needed). Thanks for any info! Tom Tom Vijlbrief TNO Institute for Perception P.O. Box 23 Phone: +31 34 63 62 77 3769 DE Soesterberg E-mail: tnosoes!tom@mcvax.cwi.nl The Netherlands or: uunet!mcvax!tnosoes!tom [[ This was discussed earlier this year. See v6n73. Also see the next message. This is important and complicated enough to warrant a full blown response. --wnl ]] ------------------------------ Date: Mon, 18 Jul 88 13:12:32 CDT From: William LeFebvre <phil@Rice.edu> Subject: login problem when NFS server crashes (2) This is such a subtle but serious problem that I will briefly explain it and the workaround here. In order to understand what is causing the problem, we need to know a little bit about how directories are scanned, how file names appear in a directory, and how "pwd" does its job. Last things first. The command "pwd" determines the name of the current directory by calling the library function "getwd". "Getwd" works by looping through the following steps until the root directory ("/") is found: stat "." to determine the inumber (or inode number); open ".." and read each entry until you find the same inumber; store the corresponding name and repeat. Unfortunately, you also have to watch for points where you cross file systems. For this reason, "getwd" stats every file it finds in ".." until it finds the right inumber. This is where the problem lies. If one of those files corresponds to an NFS file system on a downed server, the stat will hang. Guess what the c-shell does when you first start it up! Calls "getwd" to set the variable "cwd". So, if you do what most people do and put all your NFS partitions in "/", then you *might* hang when one of those servers goes down. Remember that file names do not appear in a directory in any readily discernible order. "ls" sorts them for you to make life easier on you. But they do not necessarily appear in the directory in sorted order. If you would like to see the true order in which the file names appear, use "ls -f dir". If a user's home directory is within directory X, then any NFS file system server before X in that list will cause login problems for that user when the server is down. So how do you work around the problem? In some cases, you cannot. In our particular environment, many of the system's users have all their files on the same machine. The other NFS servers contain files for various projects, the TeX library, sources for programs in /usr/local, etc. I was able to solve the problem for most of the users by making sure that the directory for their file system appeared before any other NFS file systems in "/". This is not as easy to do as I thought it would be. The positioning in a directory that a file is given is largely dependent on the file name's length. Careful playing with "mv", "rmdir" and "mkdir" should do the trick. Observe the results with "ls -f ." But you must do this with *no* file systems mounted. You can even rename a mount point if a file system is mounted there. Well, I hope this helps someone. William LeFebvre Sun-Spots moderator Department of Computer Science Rice University <phil@Rice.edu> ------------------------------ Date: Fri, 8 Jul 88 01:41:44 CDT From: mkkam@wael.cs.uh.edu (Man-Kai Kam) Subject: Level 2 interrupt problems (3/280) I configured the kernel to enable pc0 (SUN IPC) on a 3/280 which is a server. Then I remotely access it from a 3/50 client by running pctool. Right after that, the 3/280 console keeps printing "Level 2 interrupt", and the machine is so busy servicing the interrupts that it virtually halts. I can't try to run pctool locally because the 280 does not have a bitmap console (we use a vt220 as the console). Another problem is that IPC doesn't seem to like subnetting! Anyone has any hints? Please let me know. Thank you. Francis Kam Computer Science Department Internet: mkkam@sun1.cs.uh.edu University of Houston CSNET: mkkam@houston.csnet 4800 Calhoun Phone: (713)749-1748 Houston, TX 77004. ------------------------------ Date: Fri, 8 Jul 88 10:01:16 EDT From: apollo@ecf.toronto.edu (Vince Pugliese) Subject: WANTED: Expert System Software/Hardware Info I am posting this for a friend who does not have net access. Direct correspondence to his physical address is preferred though "feelers" can be sent to me and I will relay. Thanks in advance. ____________________ Our company is involved in the development of an autonomous operations planner for space station applications using AI techniques (specifically Expert Systems). We are currently investigating the purchasing of software/hardware to accomplish this task. We are looking at ES Shells as well as the possibility of using various AI languages to write our own ES. For those companies viewing this request, product information can be sent to us at the address below. We are also interested in a list of developers/manufacturers, if anyone has one. Any useful information would be appreciated. Wayne Sincarsin Dynacon Enterprises Limited 5050 Dufferin St. Suite 200 Downsview, Ontario Canada, M3H 5T5 (416) 667-0505 ------------------------------ Date: 8 Jul 88 10:57:00 EDT From: brun@nrl-lcp.arpa Subject: NCAR graphics package for the SUN? Most of the numerical codes we develop here use the NCAR graphics package to display their results. Recently we acquired a number of Sun workstations, and we are interested in learning if the NCAR package has been ported to the Sun. Has anyone run across it, and know where (if it exists) to find it? Any information would be appreciated. Todd Brun BRUN@NRL_LCP.ARPA Laboratory of Computational Physics and Fluid Dynamics Naval Research Laboratory, Washington, D.C. ------------------------------ Date: 8 Jul 88 04:16:16 GMT From: roland@wheaties.ai.mit.edu (Roland McGrath) Subject: Differences between SunOS 3.2 and 4.0? Can someone give me a list of differences between SunOS 3.2 and 4.0? It appears that 4.0 is based on 4.3 BSD and SVr3.1 rather than 4.2 BSD and SVr2.0, but I don't know if that's true or what other differences there are. If something is taken from 4.3 BSD, just "like 4.3" is sufficient, but I don't know anything about System V. Please mail responses to roland@wheaties.ai.mit.edd. Thanks. Roland McGrath roland@wheaties.ai.mit.edu, mit-eddie!wheaties.ai.mit.edu!roland [[ You're kidding, right? This is a joke, right? I think the answer would fill up a book! Doesn't Sun have a document that describes the major differences? 4.0 isn't just "based on SVr3", rather it is a peculiar merging of BSD and SV (System V). It is a product of the standardization effort that AT&T and Sun are jointly undertaking (the same effort that is being pretty much boycotted by just about all the other computer manufacturers). --wnl ]] ------------------------------ Date: Fri, 8 Jul 88 11:52:10 EDT From: dms@wheaties.ai.mit.edu (David M Siegel) Subject: backup software for Suns? We are interested in purchasing a replacement for the Sun dump program. We feel that dump is inadequate for several reasons, including: 1) tapes lack labels (would like ANSI labels) 2) single structures must go on one tape 3) poor error recovery 4) not automatic enough 5) ... I can't believe that Sun didn't rework dump under OS 4.0. If they are interested in putting together a serious operating system, the backup/restore system must be bullet-proof. These problems are exacerbated when using higher density media, like the Exabyte unit. With this device it should be possible to do an unattended incremental backup of very large filesystems given the proper dump technology. With Sun's current dump, it's all but impossible to do this reliably. It would also be nice to have file migration, but that's another story. ------------------------------ Date: Thu, 7 Jul 88 18:06:27 EDT From: Ben Shneiderman <ben@mimsy.umd.edu> Subject: UofMaryland television courses: User Interface Strategies '88 Would your circulate this notice...thanks...Ben Shneiderman ____________________ User Interface Strategies '88 A two-day national satellite TV course October 5 and 12, 1988 Organized by Ben Shneiderman, University of Maryland Presenting Thomas Malone, MIT Donald Norman, University of California, San Diego James Foley, George Washington University This course is produced by the University of Maryland Instructional Television (ITV) System and broadcast nationwide at more than 200 sites on the AMCEE/NTU (National Technological University) Satellite Network. For a copy of the full brochure and information on attending at an AMCEE site in your area or at an ITV site in the Washington, DC area, call the University of Maryland ITV office at (301) 454-8955. You may consider arranging a private showing as a special event for your organization, university, or company. Overview: New user interfaces ideas have engaged many researchers, designers, programmers, and users in the past year. These four leaders of the field offer their perspectives on why the user interface is a central focus for expanding the application of computers. Each will offer his vision and suggest exciting opportunities for next year's developments. Demonstrations, new software tools, guiding principles, emerging theories, and empirical results will be presented. Intended audience: User interface designers, programmers, software engineers, human factors specialists, managers of computer, information, and communications projects, trainers, etc. ---- October 5, 1988 ---- Ben Shneiderman, University of Maryland Lecture 1: INTRODUCTION: User Interfaces Strategies Lecture 2: HYPERTEXT: Hype or Help? Thomas W. Malone, MIT Lecture 3: COMPUTER-SUPPORTED COOPERATIVE WORK: Using information technology for coordination Lecture 4: COMPUTER-SUPPORTED COOPERATIVE WORK: Design principles and applications Discussion Hour ---- October 12, 1988 ---- Donald A. Norman, University of California, San Diego Lecture 5: USER CENTERED SYSTEM DESIGN: Emphasizing usability and understandability Lecture 6: Practical principles for designers Jim Foley, George Washington University Lecture 7: Software tools for designing and implementing user-computer interfaces Lecture 8: User Interface Management Systems (UIMSs) Discussion Hour ------------------------------ End of SUN-Spots Digest ***********************