tin@szebra.uucp (Tin Le) (02/03/91)
In article <1991Jan30.084300.13358@egsner.cirr.com> eric@egsner.cirr.com (Eric Schnoebelen) writes: > > I sat in on Dell's Sys Vr4 BOF at the Dallas Usenix last week, >and they commented that they have had no end of trouble trying to get >ttymon working on async lines. They even stated that they recommend >using getty, which is still included on all dialup async lines. Perhaps >the folks from Dell will offer us more... I've gotten ttymon to work on async lines. Even got uucp running between it and szebra (running ISC v2.0.2 SVR3.2). In this case, [it] is a Sony NEWS 3710 running S5R4 v2.0. The bugs were more likely in the async driver. I found numerous problems in the driver. > > To add a bit more meat to this posting, the Dell BOF was quite >informative. The developers from Dell were there answering questions, >and commented on the number of bugs that they found in the base line >sources from AT&T (b22, I believe). But they also claimed (and I have >no reason to doubt this) that release 4 is much more robust that release ???????????????? I disagree with that! 5.3.2 (IMHO) is more stable than current 5.4 that is being shipped out there. >3.2. Nearly all the bugs they found were in the BSD compatibility >routines. One comment (paraphrased) was "I think the people who did the >BSD stuff had never looked at a BSD machine" I agree 100% with that. We found so many bugs in the BSD emulation area. Both in the ported BSD commands/utilities and the emul libraries. <sigh>. > After attending the BOF, and chatting with the developers at >Uniforum, I would have no problems running Dell in a production >environment around here. And I plan on doing so in the near future. Be prepared for lots of problems. I have a Dell 486/33 box with their S5R4 for a week now, and have had 2 panics (and I haven't even started trying yet!). There are still many problems with what they are shipping. BTW, Dell also screwed up in their shipment. We order a complete system, but the machine arrived without the VGA monitor and the software was not completely loaded (yes, folks, UNIX was only partially on the HD). To be fair, I think the problems are also there in the S5R4 being shipped by other vendors and not just Dell. I've spent some time with as many as I could at Uniforum and my impression was they all are still in beta quality, i.e. NOT READY FOR PRODUCTION ENVIRONMENT. Anyway, it's your money and this is just IMHO. BTW, there were other problems with the system we purchased from Dell. The system booted up, but after about 10minutes, it became obvious that the various packages were only partially loaded. For example, in trying to start X, I found that many of the sym-links were not setup: /usr/X11, /usr/bin/X11, /usr/lib/X11, etc. Worse, xinit was missing, plus other rather important files/dirs: /usr/lib/X11/config, /usr/lib/X11/xinit. After some investigation, I think I understand what happened. The hard disk was a 300MB+ in size. They put both / and /usr in a single root partition, and they didn't make it large enough. So when they loaded the software onto the disk, root filled up and the install failed. Somehow this got past their test people and the system shipped to us with a partially installed system <sigh>. Hardware wise, we are quite happy with the machine. The 486/33 is definitely a screamer. It's just too bad the software business was so sloppy... -- Tin -- +----------------------------------------------------------------- |Tin Le | tin@smsc.sony.com or tin@szebra.uu.net |Station Zebra |....!{claris,limbo,sonyusa,zorch}!szebra!tin
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (02/04/91)
In article <1991Feb2.195744.15339@szebra.uucp> tin@szebra.uucp (Tin Le) writes: | Be prepared for lots of problems. I have a Dell 486/33 box with their | S5R4 for a week now, and have had 2 panics (and I haven't even started | trying yet!). I would love to know what you are doing... I had two beta versions and an alpha version of V.4 from three vendors, and never saw a panic. We had a Dell V.4 loading from tape, being an NFS server for Xenix machines, and running a scanner in direct i/o port access mode under DOS under X, all in an 8MB machine. We set a record for paging, but it all worked, reasonably fast, and nary a problem. This was on a class A network with 1k+ nodes active, so it seems fairly safe to say the net card was busy in addition to all the other stuff going on! Can you give me a hint what you are running? I have never tried a multiport serial card, although I guess I have to, since I need two serial ports and the fix for the bus mouse driver is rather seriously late. Other than that we've run for a few days on 325's, 425E, 433E with SCSI, old 16MHz SXs of no particular brand, and even a 325D. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc and 80386 mailing list "Stupidity, like virtue, is its own reward" -me
tin@smsc.sony.com (Tin "Man" Le) (02/08/91)
In article <3090@sixhub.UUCP> davidsen@sixhub.UUCP (bill davidsen) writes: >In article <1991Feb2.195744.15339@szebra.uucp> tin@szebra.uucp (Tin Le) writes: > >| Be prepared for lots of problems. I have a Dell 486/33 box with their >| S5R4 for a week now, and have had 2 panics (and I haven't even started >| trying yet!). > > I would love to know what you are doing... I had two beta versions and >an alpha version of V.4 from three vendors, and never saw a panic. We >had a Dell V.4 loading from tape, being an NFS server for Xenix >machines, and running a scanner in direct i/o port access mode under DOS >under X, all in an 8MB machine. We set a record for paging, but it all >worked, reasonably fast, and nary a problem. > I wish I could provide more details but as I noted, I was not ready at the time to test the system. The panics kinda took me by surprise so to speak. The first time was when we were experimenting with DOS Merge. We were all running it over the network to an xterm on our workstation. The second time was when we were transferring our files to the newly created accounts on the machine. Several ftp sessions were going, I believe one person had NFS mounted his home dir on the Dell box and was tar'ing his files across. There is one problem that is driving me bonker. There is a bug in the console driver (or the screen blanker). It does not matter whether I am running X or the console is just sitting at the login prompt. After some period of time (don't know how long), but the console "blank" out (i.e. go dark). From then on, it is not possible (as far as I can determine) to turn it back on except to reboot the thing. Note that the machine is still alive, i.e. I can rlogin to it and use it. Just the console display is GONE. The funny thing is, I can actually login at the console (if it was sitting at the login prompt), but I just can't see what I am doing. This bug is very consistent and is bad enough that we just got tired of rebooting it. Since the machine sits in a back room and most of us just rlogin to it anyway, we have not done much about it. By the way, I've tried various things to try bringing the console back. Some of these are, logging on at the console and trying to run X (hoping that will bring the screen alive...no such luck), or running 'reset', 'tset', etc... -- Tin -- .---------------------------------------------------------------------- . Tin Le Work Internet: tin@smsc.Sony.COM . Sony Microsystems UUCP: {uunet,mips}!sonyusa!tin . Work: (408) 944-4157 Home Internet: tin@szebra.uu.net
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (02/11/91)
In article <1991Feb7.222422.19476@smsc.sony.com> tin@smsc.sony.com (Tin "Man" Le) writes: | There is one problem that is driving me bonker. There is a bug in | the console driver (or the screen blanker). It does not matter whether | I am running X or the console is just sitting at the login prompt. | After some period of time (don't know how long), but the console | "blank" out (i.e. go dark). I have no ideas on that one, my X blanks nicely and comes back as expected, which the regular console never blanks. In fact I wrote a little command which does the blank, and I manually type blank when I'm going away. I don't feel ready to write a good blanker yet, and I don't want a hack. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc and 80386 mailing list "Stupidity, like virtue, is its own reward" -me
mtymp02@ux.acs.umn.edu (John T. Alexander) (02/15/91)
If you want to see some major increased system performance under Unix, partition the drive so that root,usr have their own file systems. You think a 486/33 is a screamer. Kick in an EISA bus with a Mylex 1 Meg cache controller.
prewitt@convex.com (Mike Prewitt) (02/21/91)
In article <3278@ux.acs.umn.edu> mtymp02@ux.acs.umn.edu (John T. Alexander) writes: >If you want to see some major increased system performance under Unix, >partition the drive so that root,usr have their own file systems. > >You think a 486/33 is a screamer. Kick in an EISA bus with a Mylex 1 Meg >cache controller. Isn't that what Berkeley did so long ago??? ( and they even made a file sytem to do it too :-) -- Michael Prewitt Software Test Design Engineer Convex Computer Corperation prewitt@convex.com P.O. Box 833851 Richardson Texas, 75083-3851
prewitt@convex.com (Mike Prewitt) (02/21/91)
In article <3278@ux.acs.umn.edu> mtymp02@ux.acs.umn.edu (John T. Alexander) writes: >If you want to see some major increased system performance under Unix, >partition the drive so that root,usr have their own file systems. Didn't Berkeley do this a long time ago??? :-) Michael Prewitt Software Test Design Engineer Convex Computer Corperation prewitt@convex.com P.O. Box 833851 Richardson Texas, 75083-3851 -- Michael Prewitt Software Test Design Engineer Convex Computer Corperation prewitt@convex.com P.O. Box 833851 Richardson Texas, 75083-3851
richard@pegasus.com (Richard Foulk) (02/22/91)
>You think a 486/33 is a screamer. Kick in an EISA bus with a Mylex 1 Meg >cache controller. EISA is unnecessary. ISA with the Mylex 1 (or 4) meg cache controller is pretty fine itself.