Sun-Spots-Request@RICE.EDU (William LeFebvre) (06/14/88)
SUN-SPOTS DIGEST Monday, 13 June 1988 Volume 6 : Issue 110 Today's Topics: Re: LaserWriter IIs on Suns Re: X11 & Suntools on a Sun 3/50 Re: slip on Suns Recommended books, TCP/IP and Ethernet Philips monitors release 4.0 and file system performance Problems with a DELNI on a Sun? panic 'iechkcca' freeware draw program for Suns? what to do after clear_colormap? modifying 4.0 library? Info on Large Screens? wrong shading colors in suncore? 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: Mon, 6 Jun 88 13:57:00 CDT From: sob@watson.bcm.tmc.edu (Stan Barber) Subject: Re: LaserWriter IIs on Suns We have been running our LaserWriter IIs on our Suns for about three months. We are running Transcript 2.1. We have had no problems. I guess we are doing something right. Stan internet: sob@tmc.edu Baylor College of Medicine Olan uucp: {rice,killer,hoptoad}!academ!sob Barber Opinions expressed are only mine. ------------------------------ Date: Mon, 6 Jun 88 16:53:21 EDT From: Neil Movold - DCS User Services <movold@watdcsu.waterloo.edu> Subject: Re: X11 & Suntools on a Sun 3/50 In v6n85, Isaac Salzman describes a method of creating multiple screens using X11 and Sunview. The idea appeals to us here at Waterloo as our Suns do not support dvitool but do support xdvi for previewing dvi output. I have run through the instructions given by Isaac, but I have run into some problems : 1) Our 3/50 workstations only ( from what I can tell ) support bwtwo0 frame buffer in /dev which when used allows output from the primary window to over write onto the secondary window. This is alright except for continuously running tools such as perfmeter and clock or output returned from background jobs. Can bwtwo1 be generated on a 3/50 ??? 2) The other major problem I have is that when I xinit Xwindows using overview, I loose keyboard control in Suntools. I have managed to put a menu item in my .rootmenu file that allows me to reset the keyboard in Suntools, but when I do that I loose keyboard control in Xwindows. Has anyone run into this problem, and what possible action can I take to solve the problem ??? As we do not have many 3/60c, I would like to try and get this thing running on our 3/50s. Any suggestions would be appreciated.... Neil Movold Department of Computing Services University of Waterloo Waterloo, Ontario, Canada <movold@watdcsu.UWaterloo.ca> ------------------------------ Date: Mon, 6 Jun 88 22:57:55 MDT From: Mark G. Brown <markb@boulder.colorado.edu> Subject: Re: slip on Suns >From: Ned Danieley <ndd@sunbar.mc.duke.edu> > >I have installed the slip code that I obtained from the Sun-spots >archives, and have run into a couple of problems. sunbar is our server, >and is on our LAN; bme is a remote machine that is slipped to sunbar. >sunbar and bme seem to communicate just fine (rlogin, rcp, ping), but bme >can't talk to any of the other machines on our LAN (like dukebar).... our situation sounds similar. we have a sub-net of suns off-campus served by wasser, which is connected to sigi on campus via slip, and their internet addresses are: 128.138.240.4 sigi # vax785, 4.3BSD 128.138.240.2 wasser-gw # sun3/280, SunOS 3.5 128.138.239.1 wasser # sun3/280, SunOS 3.5 we had to treat the slip line as a gateway, and place it on the same subnet as sigi. here are the corresponding entries in wasser:/etc/rc.local: /etc/slattach /dev/ttya 9600 /etc/dstaddr sl0 128.138.240.4 /etc/ifconfig sl0 128.138.240.2 up arp -trailers netmask 0xffffff00 /usr/etc/route add 128.138.240.2 localhost 0 /usr/etc/route add default 128.138.240.2 10 and sigi:/etc/rc.local: /etc/slattach ttyB1 9600 >/dev/console ifconfig sl0 inet 128.138.240.4 128.138.239.1 up subarp\ netmask 0xffffff00 >/dev/console hope this helps. -markb Mark G. Brown, University of Colorado, Boulder, Colorado 80309 (303)-492-3972 markb@sigi.colorado.edu -or- !{nbires|ncar}!boulder!markb ------------------------------ Date: 6 Jun 1988 1852-PDT (Monday) From: Charles Spurgeon <spurgeon@jessica.stanford.edu> Subject: Recommended books, TCP/IP and Ethernet The following is a list of books I wrote up recently. It's meant to help someone interested in finding material on TCP/IP protocols and Ethernets. Since that is a task a lot of Sun workstation administrators are faced with, I thought I'd send the list along. The list is longish (about 2K words) since I included all the access information I could find, as well as some annotations. [[ Well, it's a little longer than that: 13347 bytes. It has been placed in the archives as "sun-spots/Ethernet.books". It can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server with the request "send sun-spots Ethernet.books". For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] Cheers, Charles Spurgeon spurgeon@jessica.stanford.edu Stanford University Networking and Communications Systems Dept. June 6, 1988 ------------------------------ Date: Tue, 7 Jun 88 11:15:22 CDT From: gene@uicsl.csl.uiuc.edu (Gene Gardner) Subject: Philips monitors Although this information has been offered in previous scattered messages, the failure-trend is now clearly established so that SUN (Philips) monitor users can save $100's if not $1000's of unnecessary repair-bills by taking these suggestions seriously. Please refer to SUNSPOTS msg v6n100 (Philips monitors-JUST SAY NO!), and v6n102 (Re: SUN monitors hardware problem). Over the last two or three years...about 10% (15 out of 150 units in our building) have failed with one, or both of these specific component failures: i.e. the "flyback transformer", or C-209. Either fault can be easily and quickly repaired in less than an hour...even at the user-site....and is so arbitrary that a relatively inexperienced technician can do it. Note: just in the last two days, I have had two more "C-209's" open up... one of which was a recent unit (different case, on its own swivel-base). Just a brief review: The "flyback" is now available from SUN, or from Dyn Service Network in San Jose, CA ...for about $65 to $85, and can be removed through the top cover (only two screws, and a plug-in connector), and C-209 is on the right-hand PC-board as you face the screen....it can be loosened and tilted without removing any connectors. C-209 is now an electrolytic (questionable choice for this application) 10uf-50v....but it should be replaced with a tantulum. I used 10uf-35v (only 12vdc across it) and can be tack-soldered on the underside of the PC board. BE SURE to observe POLARITY. It is permissible to leave the original in place if you like. It may sound a little presumptious, but I would suggest that if you screen all of your faulty Philips monochrome monitors for these two defects, you will avoid at least half of your expense, not to mention inconvenience and 30-days or more lost-time. THE PITY OF IT IS, that these appear to be very nice monitors and all the the ill-will such as described in SUNSPOTS v6n100 could be avoided if SUN simply stopped stone-walling and adopted a cooperative policy of offering schematics on the monitor-portion....or at a minimum, offering individual sub-unit repair, or exchange at reasonable prices. We also have several of their color monitors which are beatifully designed for quick-easy sub-unit exchange. We recently had a failure and were able to quickly determine which sub-unit was bad by interchanging with a good unit. What a shame that SUN has no exchange, or repair alternative on such sub-units(much grief and BIG BUCKS would be saved). [[ Mr. Gardner later sent me this addendum: --wnl ]] I have had a single failure (so it certainly can't be considered a failure-trend from our viewpoint) that gives almost exactly the same symptoms as "open C-209" (Philips monitors): It is a faulty C-207, (located near C-209). But it must be removed, and should be replaced with a good-quality (temperature-stable) 1000pf (.001ufd) capacitor. ------------------------------ Date: Mon, 6 Jun 88 21:23:42 EDT From: dan@wind.bellcore.com (Daniel R. Strick) Subject: release 4.0 and file system performance Cpu and possibly other overhead for file system access is way up under release 4.0 (compared to release 3.x). I just measured a 36% increase in cpu for reads and a 16% increase for writes (on a sun-3/260 with 8mb main memory). I don't know how much the increase is likely to vary with the phase of the moon or what effect this will have on normal timesharing, but the increase is quite noticeable in disk benchmarks. Peak file system i/o rates are down from about 1500 kb/sec to about 1200 kb/sec. The exact source of the new overhead is not clear, but the use of virtual memory pages for buffers is implicated. Heavy file system i/o consumes essentially all of free memory, drives the virtual memory scan rate up to something between 100 and 200, and seems to cause a little paging and swapping. I don't think this is really the problem. The increase in memory available for the buffer pool is a big win, but this also creates an opportunity to spend more cpu browsing through buffer and page lists (32 mb used to get you 400 buffers; now it gets you about 3500). Lots of real memory is likely to reduce disk i/o, but it also encourages i/o patterns easily optimized by smart disk controllers/drivers and it seems a shame to lose 20% of your peak i/o rates to kernel overhead. Does anyone think they know exactly where release 4.0 eats the extra cpu cycles? Processor cache management for DVMA is a cpu pig, but this is not new in release 4.0. Cpu overhead for disk i/o on 200 series suns is now approaching 10 times what it is on a vax 8650. I would like to see this trend reversed. ------------------------------ Date: Tue, 07 Jun 88 09:58:28 -0400 From: lrj@helios.tn.cornell.edu Subject: Problems with a DELNI on a Sun? Reference: v6n104, v6n96 > ... Sun says that this is normally caused by > either broken hardware or some problem with Ethernet cabling (e.g. use of > a DEC DELNI). ... What problems are there with DELNI's? I've been running my network for two years now with a DELNI as an integral part. The current configuration is 1 DEC DELNI, 3 Cabletron MT-800's (another multiport tranceiver), ~20 Suns and a Proteon P2400 gateway. This has been working fine for quite some time. If there are known problems with DELNI's & Suns, what should i look for? -- Lewis R. Jansen, LASSP Systems Grunt lrj@helios.tn.cornell.edu ------------------------------ Date: Mon, 6 Jun 88 20:19:29 PDT From: Stuart Cracraft <cracraft@hyper-sun1.jpl.nasa.gov> Subject: panic 'iechkcca' There seems to be a recurring panic, known about by SUN, that is fairly subtle -- the symptom is that NFS fails with a 'getattr' error and within a short period of time thereafter, one or more SUN's that remote-mount a filesystem from the SUN about which the 'getattr' was reported will panic with the 'iechkcca'. All I have gotten so far from SUN is that this may be related, somehow, to the Xylogics disk controller on the SUN exporting the NFS filesystem. Has anyone else observed this? This is a fairly irksome occurrence because it can have wide implications on local networks. Can SUN comment further on this and what their plans are to provide a work-around or preferably a fix? Stuart [[ Charles Hedrick says that an iechkcca panic happens when the Intel Ethernet controller chip does not acknowledge a command after a reasonable amount of time: either because of broken hardware or an Ethernet cabling problem. He also said that he believes it can happen when all the hardware is fine. He also recommends calling the support group at 800-USA-4SUN. See Sun-Spots digest v6n104 for hsi recent posting on this subject. --wnl ]] ------------------------------ Date: Mon, 6 Jun 88 15:27:22 PDT From: celeste@coherent.com (Celeste C. Stokely) Subject: freeware draw program for Suns? Does anyone have a free drawing program for Suns, under at least 3.4, under Sunview, able to input and output Sun raster files? Something like MacPaint functionality would be fabulous. The "gwpaint" on the '85 SUG tape isn't exactly robust. Ok to tell me to RTFArchives if there's a paint program waiting for me there. ..Celeste Stokely Coherent Thought Inc. UUCP: ...!{ames,sun,uunet}!coherent!celeste Domain: celeste@coherent.com Internet: coherent!celeste@ames.arpa or ...@sun.com or ...@uunet.uu.net VOX: 415-493-8805 SNAIL:3350 W. Bayshore Rd. #205, Palo Alto CA 94303 [[ The archives contains two programs that might interest you. A M*cDraw-like program called "fig" and a M*cPaint-like program called "touchup". Fig is in nine shar files, named fig.shar.0X (for 1<=X<=9) and touchup is in 6 shar files, named touchup.shar.X (for 1<=X<=6). They are all in the "sun-source" directory of the archives and can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server. For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] ------------------------------ Date: 7 Jun 88 01:45:45 GMT From: mkhaw@teknowledge-vaxc.arpa (Mike Khaw) Subject: what to do after clear_colormap? Is there any graceful way to recover after a "clear_colormap" (SunOS 3.5) turns the screen blank white -- if you were running suntools and now have no mouse pointer? Thanks, Mike Khaw internet: mkhaw@teknowledge.arpa uucp: {uunet|sun|ucbvax|decwrl|uw-beaver}!mkhaw%teknowledge-vaxc.arpa hardcopy: Teknowledge Inc, 1850 Embarcadero Rd, POB 10119, Palo Alto, CA 94303 ------------------------------ Date: 7 Jun 88 02:58:47 GMT From: hedrick@athos.rutgers.edu (Charles Hedrick) Subject: modifying 4.0 library? We are trying to bring up 4.0 this summer. Unfortunately, we don't have source yet, and based on past experience, we have to assume it's going to be a while before we can get it. One very high priority change is to modify gethostbyname and friends so it uses the domain resolver instead of yp. In the past it was easy to replace a few routines in libc.a: ar could simply be told to replace them. However now we've got this libc.so thing. Anybody - know how to replace a couple of modules in such a file - know how to build libc.so from libc.a - have a copy of libc.so with the resolver routines instead of the usual yp junk? I know there is a libresolv, but since we don't have source to the utilities either, we can't relink the existing system software to use it. So the only hope I can think of is to fix the sharable library and hope that all the system software is linked -Bdynamic. ------------------------------ Date: Tue, 7 Jun 88 08:38:40 PDT From: smith@phylo.life.uiuc.edu (Steven Smith) Subject: Info on Large Screens? I'm currently working with the University of Illinois Department of Microbiology, and we are in the process of designing a large scale DNA Sequence editor/analysis package. Our specifications require that we be able to display huge amounts of text at one time (on the order of 300x100 characters). Although our current displays can come close to this level, it is accomplished by means of using smaller fonts. This method is not very practical for long term use because of eye strain. Does anyone know of any third party companies that sell larger displays for the sun? 2000x2000 resolution would be optimal, and color or grey scale prefered. Any responses are welcome, company addresses and/or phone numbers prefered. Thank you, Steven Smith U of Illinois, Microbiology smith@phylo.life.uiuc.edu ssmith@ncsavmsa.bitnet ------------------------------ Date: Tue, 07 Jun 88 17:17:47 MEZ From: TS36@DFVLRGO1.BITNET Subject: wrong shading colors in suncore? Using the SUN-CORE graphics a lot of mysterious this happen to me. One is very essential: I use the SetVertexIndices routine to define the color of a polygon. Then I draw a PolygonAbs3 using the Gouraud shade. Now rotate that thing with SetSegImgXform3 and it appears in arbitrary color. Somehow the color seems to depend on the angle of rotation. all this happens when I initialize the cgpixwindd. if i run the same program but initialize gp1pixwindd everything is perfect and all colors appear as they should do. OK, I really don't expect to find anybody who knows a solution to this problem. Even so I would very much appreachiate if I could. But as people at SUN in Germany tend to tell me that I am the only one in the world with such strange encounters, I would be glad to get notes from people who have the same or similar or even different problems with SUN-CORE May be one day we are enough people to encourage SUN to fix the problem. hoping for e-mail schorsch pagendarm dfvlr bunsenstr.10 3400 goettingen federal republic of germany TS36@dfvlrgo1.bitnet ------------------------------ End of SUN-Spots Digest ***********************