iaf@ccadfa.adfa.oz.au (- Mathematics) (03/06/91)
Why is there always major problems upgrading the operating system? Why do other users in the USA say they have had no problems upgrading to sr10.3? (previous news) Two major problems have been encountered - 1) The upgrade somehow managed to whipe our registry clean except for 5 accounts!! This process was also repeated at another site on campus with the same results. Question: Are all the other sites which upgraded and had no problem open or closed sites? Question: Did we get a bum distribution? The problem took 4 programmer days to restore, with no answers from our local Australian Support office. 2) sr10.3 and lpd do not mix. Our printers require full XON/XOFF so they can tell the apollo when their buffer is full. So they need to have -insync (tctl command) BUT each time a file is sent to any printer on any sio line insync is reset to FALSE thus some pages become lost... Question: Does anyone out there use lpd and have printers directly connected to apollo sio lines? Question: Do most apollo users only use AEGIS and not Unix? Question: Does anyone have a solution? Thanks in advance, Annoyed and frustrated system administrator, Ian. ################################################################################### Ian French # Phone : +61 6 2688882 Mathematics Department, # Fascimile : +61 6 2688886 University College, # E- Mail : (INTERNET) iaf@maadfa.ma.adfa.oz.au University of New South Wales, # Australian Defence Force Academy, # "Upward compatibility is a ruse to keep always Northcott Drive, CANBERRA ACT # on the move, always upgrading, always buying Australia 2600 # and always busy" ###################################################################################
etb@milton.u.washington.edu (Eric Bushnell) (03/07/91)
In article <2300@ccadfa.adfa.oz.au> iaf@ccadfa.adfa.oz.au (- Mathematics) writes: > > Question: Are all the other sites which upgraded and had no > problem open or closed sites? The 10.2's were mixed--open and closed. All 10.3 installations and upgrades have been closed--such as it is. (see earlier news threads) > > Question: Does anyone out there use lpd and have printers > directly connected to apollo sio lines? I have a laserjet 3 connected as plain line printer (waiting for a PostScript cartridge) driven by lpd. I use a basic 4-wire serial cable, and the plain lp printcap entry. This is only one node, though. The other printers on other networks run from the Aegis print servers, because they were set up that way when I got here and I haven't had reason to change that. I have had trouble getting lpd to work on the same node that the Aegis print software runs on, but I haven't played with that combination enough to know how serious a problem it is. > > Question: Do most apollo users only use AEGIS and not Unix? I loaded Aegis largely because some of my users expect it--it's all they've learned. Since this is a university, however, the user community is changing, and I'm moving them toward BSD Unix. I prefer BSD because it's a familiar face to university computing folk. > Question: Does anyone have a solution? Sacrifice a virgin. Say the magic words "A la peanut butter sandwich." Get a bigger hammer. All seriousness aside...I do tend to rely on the Aegis print manager/server software rather than lpd. It seems to be easier to configure and run than lpd. Of course, it only supports a few printers, and you're probably lucky enough not to have one of them. Eric Bushnell UW Civil Engr etb@zeus.ce.washington.edu
thompson@PAN.SSEC.HONEYWELL.COM (John Thompson) (03/07/91)
> Why is there always major problems upgrading the operating > system? > Why do other users in the USA say they have had no problems > upgrading to sr10.3? (previous news) > > Two major problems have been encountered - > > 1) The upgrade somehow managed to whipe our registry > clean except for 5 accounts!! This process was also repeated at > another site on campus with the same results. > Question: Are all the other sites which upgraded and had no > problem open or closed sites? We installed as 'open' and then locked it up w/ a hand-made inprot template > Question: Did we get a bum distribution? > The problem took 4 programmer days to restore, with no answers > from our local Australian Support office. I'd say that you either got a bad distribution, or that you (someone else?) has modified some of the installation, perhaps by using scripts that do the actual install. If somebody put in / left in a call to crrgy, that would certainly screw things up. Another possibility: did you boot off the cartridge tape provided? We didn't (and you certainly don't have to), so there might be a problem if you start up that way. > 2) sr10.3 and lpd do not mix. Our printers require full > XON/XOFF so they can tell the apollo when their buffer is full. > So they need to have -insync (tctl command) BUT each time > a file is sent to any printer on any sio line insync is reset > to FALSE thus some pages become lost... > Question: Does anyone out there use lpd and have printers > directly connected to apollo sio lines? 'fraid not (yet). > Question: Do most apollo users only use AEGIS and not Unix? We use primarily Aegis. However, even mor than that, we use the Mentor print servers, which act a lot like pre-sr10 Aegis servers. I think that a good chunk of the sites use Aegis, and a good chunk use Unix. Apollo has traditionally given more support to Aegis (my personal opinion). That's why it hurts so bad now that they're going away from it. -- jt -- John Thompson Honeywell, SSEC Plymouth, MN 55441 thompson@pan.ssec.honeywell.com Me? Represent Honeywell? You've GOT to be kidding!!!
etb@milton.u.washington.edu (Eric Bushnell) (03/07/91)
>In article <2300@ccadfa.adfa.oz.au> iaf@ccadfa.adfa.oz.au (- Mathematics) writes: >> >> Question: Does anyone out there use lpd and have printers >> directly connected to apollo sio lines? > >I have a laserjet 3 connected as plain line printer (waiting for >a PostScript cartridge) driven by lpd. I use a basic 4-wire serial >cable, and the plain lp printcap entry. This is only one node, > >I have had trouble getting lpd to work on the same node that the >Aegis print software runs on, but I haven't played with that >combination enough to know how serious a problem it is. >> Since then, I have tried getting lpd to run on my token ring, which has a 10.3 node as print manager and server. Sure enough, it didn't work. The prsvr locks /dev/sio1, so a printcap spec of lp=/dev/sio1 on that host reports device busy. (Makes sense) I tried running the lpd on another host, with a printer on the sio1 port. Then lpr reports "no daemon present." Apparently the master lpd won't fork properly. Curiouser and curiouser. The ethernetworked host that runs lpd and a laserjet w/o trouble is down today. Strange that it should work so easily, given these troubles. Is the token ring doing something strange? I remember some discussion of lpd here in recent weeks. Anybody wanna repost a summary? Sounds like a subject for the FAQ file, too. > >> Question: Does anyone have a solution? > >Sacrifice a virgin. Say the magic words "A la peanut butter >sandwich." Get a bigger hammer. > Time to find a REALLY big hammer... > Eric Bushnell UW Civil Engr etb@zeus.ce.washington.edu
krowitz@RICHTER.MIT.EDU (David Krowitz) (03/07/91)
There seems to be some (more) bugs in the SIO line support at SR10.3 Apparently even the DTR/CTS flow control does not work for some CPU models -- the sense of the lines (high vs. low voltage) was accidentally reversed! If I remember correctly the person who found this little "feature" was running on a HP/Apollo 9000 model 400s. If people feel like sending me their SR10.3 bugs (and work-arounds!) I'll compile a summary for the network. == Dave
chen@digital.sps.mot.com (Jinfu Chen) (03/08/91)
In article <2300@ccadfa.adfa.oz.au> iaf@ccadfa.adfa.oz.au (- Mathematics) writes: > > 1) The upgrade somehow managed to whipe our registry > clean except for 5 accounts!! This process was also repeated at > another site on campus with the same results. > Question: Are all the other sites which upgraded and had no > problem open or closed sites? > Question: Did we get a bum distribution? > The problem took 4 programmer days to restore, with no answers > from our local Australian Support office. Are you upgrading from 10.1 to 10.3? Like someone pointed out before, the 10.2/10.3 registry server is not compatible with 10.1. We were lucky by going to 10.2 rgyd long before 10.3 shows up. However, the 10.3 release notes doesn't mention this incompatibility at all. -- Jinfu Chen (602)898-5338 Motorola, Inc. SPS Mesa, AZ ...uunet!motsps!digital!chen chen@digital.sps.mot.com CMS: RXFR30 at MESAVM ----------
fridman@cpsc.ucalgary.ca (fridman) (03/08/91)
I'm trying to hook up a printer to one of the serial lines on a DN4500 running SR10.3. What is the difference between /dev/sio1, /dev/tty01 and /dev/tty1. Furthermore, how can I set XON/XOFF using the stty comand? ( stty tandem > /dev/[sio1,tty01] did not enable software flow control but it did set the flag. ) I know that tctl -line 1 insync works great, BUT HOW DO I DO THE SAME THING USING BSD COMMANDS?????? RF.
fridman@aa.cpsc.ucalgary.ca (Robert Fridman) (03/08/91)
Ian French writes:
2) sr10.3 and lpd do not mix. Our printers require full
XON/XOFF so they can tell the apollo when their buffer is full.
So they need to have -insync (tctl command) BUT each time
a file is sent to any printer on any sio line insync is reset
to FALSE thus some pages become lost...
Question: Does anyone out there use lpd and have printers
directly connected to apollo sio lines?
I'm running into the same problem.
My setup:
10.3, DN4500, apple laserwriter
I'm using strictly BSD printing with Adobe TransCript. When printing small
files, all is well. But as soon as a larger file is spooled, I get the
error %%[ ioerr ...]%% from the printer in the error file in the
/usr/spool/lpd/<printer> directory.
This error means that the buffer has been overrun. After hooking up a
serial analizer, it was confirmed that the printer was sending ^S but
the apollo was ignoring it.
The stty command has a tandem flag which, according to the man page is supposed
to work XON/XOFF but is doesn't.
Using tctl -insync works in setting XON/XOFF, but where are you going to put
this command? (I tried putting /com/tctl ... into the psif script but it
has no effect.)
Either stty tandem is broken or I'm misinterpreting it!!!!!!!!
RF.
etb@milton.u.washington.edu (Eric Bushnell) (03/08/91)
In article <17856@milton.u.washington.edu> etb@milton.u.washington.edu (Eric Bushnell) writes: >Curiouser and curiouser. The ethernetworked host that runs >lpd and a laserjet w/o trouble is down today. Strange that >it should work so easily, given these troubles. Is the >token ring doing something strange? No, I'm just not thinking straight. Mea culpa. The node that runs lpd is still at 10.2 after all. Sorry about the red herring. Has anybody reported this to apollo? Eric Bushnell UW Civil Engr etb@zeus.ce.washington.edu
rees@pisa.citi.umich.edu (Jim Rees) (03/09/91)
In article <1991Mar7.221511.28836@cpsc.ucalgary.ca>, fridman@aa.cpsc.ucalgary.ca (Robert Fridman) writes:
Either stty tandem is broken or I'm misinterpreting it!!!!!!!!
You're misinterpreting it!!!!!!!! Tandem mode controls xon/xoff on output.
What you want for your printer is xon/xoff on input, which is controlled by
the start and stop characters!!!!!!!!
obrennan@CC3.CC.UMR.EDU (obrennan) (03/09/91)
I'm trying to hook up a printer to one of the serial lines on a DN4500 running SR10.3. What is the difference between /dev/sio1, /dev/tty01 and /dev/tty1. Furthermore, how can I set XON/XOFF using the stty comand? ( stty tandem > /dev/[sio1,tty01] did not enable software flow control but it did set the flag. ) I know that tctl -line 1 insync works great, BUT HOW DO I DO THE SAME THING USING BSD COMMANDS?????? RF. --------------------------------------------------------- In the printcap file you can use XC and XS to clear and set the LOCAL MODE bits, and FC and FS to clear and set the SG_FLAGS as documented in the MAN pages section 4 for TTY. These flag bits control the tty interface. The ^Q/^S control mode (same as -insynch) is set by the LOCAL MODE bits. Hope this helps more than it confuses... Gerry O'Brennan Programmer/Analyst II Computing Services University of Missouri - Rolla ------------------------------ obrennan@apollo.cc.umr.edu c0022@umrvmb.umr.edu ------------------------------
fridman@cpsc.ucalgary.ca (fridman) (03/15/91)
In article <1991Mar7.221511.28836@cpsc.ucalgary.ca>, fridman@aa.cpsc.ucalgary.ca (Robert Fridman) writes:
Either stty tandem is broken or I'm misinterpreting it!!!!!!!!
You're misinterpreting it!!!!!!!! Tandem mode controls xon/xoff on output.
What you want for your printer is xon/xoff on input, which is controlled by
the start and stop characters!!!!!!!!
But I do want it on output since its the apollo that is outputting
and its the apollo that is not honorig ^S/^Q.
Besides, tandem does not work!!!!!!
RF.