dosa@delta.eecs.nwu.edu (Chonghoon Kim) (02/26/91)
I am trying to install X11R4 on my DN2500 station. I ftp'd core distribution from mordred.cs.purdue.edu. After compilation and installation, I found that the server (Xapollo) won't communicate with client programs. Can anybody comment about this disaster? I'll really appreciate it. Thanks. dosa@delta.eecs.nwu.edu
mehlhaff@hailstorm.Berkeley.EDU (Eric Mehlhaff) (02/28/91)
In article <16882@accuvax.nwu.edu> dosa@delta.eecs.nwu.edu (Chonghoon Kim) writes: >I am trying to install X11R4 on my DN2500 station. I ftp'd >core distribution from mordred.cs.purdue.edu. After compilation >and installation, I found that the server (Xapollo) won't >communicate with client programs. Can anybody comment about >this disaster? I'll really appreciate it. Thanks. > I've had similar problems here. And some new ones. Basically, the clients not communicating with the server seems to be a random thing, and doesn't happen all the time. Maybe 1 in 20 can't connect. But if that one is the one at the end of your Xsession, well, then you get logged out... AS far as new ones, on our diskless nodes, when someone logs out of a session, the xdm seems to have problems resetting, and we end up with a blank (grey) screen with no login window. APollo's X that came with 10.3 did this too. We thought MIT X11R4 would fix this, but no luck. And even more serious, our X server occasionally dies, and won't come back up. When we try to restart it, the gpr_$init() call fails to set the screen and the X server gives up. Sometimes it's aws_$init that fails instead. And sometimes the server comes up but refuses to accept any keyboard input. I downloaded our X source from gatekeeper.dec.com, put in all 18 patches, and compiled it with cc 6.8, under 10.3. A version compiled with cc 6.7 had the same problems. I'm not sure where to go from here... ERic Mehlhaff OCF stafF
rees@pisa.citi.umich.edu (Jim Rees) (03/03/91)
In article <1991Feb28.051945.20170@agate.berkeley.edu>, mehlhaff@hailstorm.Berkeley.EDU (Eric Mehlhaff) writes:
I've had similar problems here. And some new ones. Basically, the
clients not communicating with the server seems to be a random thing, and
doesn't happen all the time. Maybe 1 in 20 can't connect. But if that one is
the one at the end of your Xsession, well, then you get logged out...
I don't know if this is your problem, but...
You shouldn't try to use UDS for X. UDS has bugs related to the size of
messages you pass through it. Besides which, it's actually slower than TCP.
Turn on your loopback interface ("lo0") and use that instead. Either put
"localhost" in /etc/hosts or set your display to the dotted-decimal address
so you don't have to go through name service every time you start up an X
client.mehlhaff@whirlwind.Berkeley.EDU (Eric Mehlhaff) (03/03/91)
In article <502114d0.1bc5b@pisa.citi.umich.edu> rees@citi.umich.edu (Jim Rees) writes: >In article <1991Feb28.051945.20170@agate.berkeley.edu>, mehlhaff@hailstorm.Berkeley.EDU (Eric Mehlhaff) writes: >> >> I've had similar problems here. And some new ones. Basically, the >> clients not communicating with the server seems to be a random thing, and >> doesn't happen all the time. Maybe 1 in 20 can't connect. But if that one is >> the one at the end of your Xsession, well, then you get logged out... > >I don't know if this is your problem, but... > >You shouldn't try to use UDS for X. UDS has bugs related to the size of >messages you pass through it. Besides which, it's actually slower than TCP. > >Turn on your loopback interface ("lo0") and use that instead. Either put >"localhost" in /etc/hosts or set your display to the dotted-decimal address >so you don't have to go through name service every time you start up an X >client. I have been setting the dotted-decimal addresss! Here's a sample from a random xdm-errors file: XIO: fatal IO error 32 (Broken pipe) on X server ":0.0" after 45 requests (37 known processed) with 0 events remaining. The connection was probably broken by a server shutdown or KillClient. XIO: fatal IO error 32 (Broken pipe) on X server ":0.0" after 6782 requests (6780 known processed) with 3 events remaining. The connection was probably broken by a server shutdown or KillClient. Other than these errors, I'm getting a lot of Invalid Argument errors, but I I think that's something different: XIO: fatal IO error 22 (Invalid argument) on X server ":0.0" after 201 requests (194 known processed) with 0 events remaining. Here's another bug that's made itself really annoying: Periodically, xclocks and oclocks will seem to take over the machine. Unsuspecting users who are running these programs find their machines locking up every once in a while, and the only way of freeing the machine is rsh-ing in and killing the clock. I saw mention of a clock bug on apollos here not too long ago. Is this related? ------------------------------------------------------------------------------------------- ERic Mehlhaff printer-czar@ocf, mehlhaff@ocf.Berkeley.EDU The OCF -- BAD computers for BAAD people
tomg@hpcvlx.cv.hp.com (Thomas J. Gilg) (03/08/91)
> You shouldn't try to use UDS for X. UDS has bugs related to the size of > messages you pass through it. Besides which, it's actually slower than TCP. Under SR10.3, UDS is faster. Prior to SR10.3, UDS was indeed slower than TCP. In the pskq2_91 release of the borrow mode server /etc/Xdomain, we worked around the writev() problem (sum of iovs > 128k). Thomas Gilg tomg@cv.hp.com
hanche@imf.unit.no (Harald Hanche-Olsen) (03/09/91)
> You shouldn't try to use UDS for X. UDS has bugs related to the size of > messages you pass through it. Besides which, it's actually slower than TCP. Under SR10.3, UDS is faster. Prior to SR10.3, UDS was indeed slower than TCP. UDS? What's that? (Do I need to know?) % man -k uds uds: nothing appropriate - Harald Hanche-Olsen <hanche@imf.unit.no> Division of Mathematical Sciences The Norwegian Institute of Technology N-7034 Trondheim, NORWAY
rand@HWCAE.CFSAT.HONEYWELL.COM (03/12/91)
>>>>> Thomas Gilg <tomg@cv.hp.com> on 7 Mar 91 21:51:14 GMT in [Installation of X11R4] writes: > [...] > In the pskq2_91 release of the borrow mode server /etc/Xdomain, we worked > around the writev() problem (sum of iovs > 128k). Ok. So where do I get this PSK? Is it released yet? -- Douglas Keenan Rand Honeywell -- Air Transport Systems Division Phone: +1 602 436 2814 US Snail: P.O. Box 21111 Phoenix AZ 85036 Internet: @cim-vax.honeywell.com:rand@hwcae.cfsat.honeywell.com -or- drand@hi-csc.honeywell.com UUCP: ...!uunet!asuvax!apciphx!hwcae!rand
tomg@HPCVXPH.CV.HP.COM (Thomas J. Gilg) (03/12/91)
pskq2_91 is a formal Domain/OS release available from Apollo to be installed on top of SR10.3 Contact your Apollo sales rep I suppose. Up front, I'm an engineer and haven't a clue how customers get things :-( If you can't track something down, let me know, and I'll see what I can do. The manufacturing part number is 018890-A00 "Quarterly PSK Q2-91" If you want X11R4 support AND 68040 support, there is souped up version of psk8 (68040 support) with X11R4 added. The manfacturing part number is 019295-A00 "PSK8X_040" Thomas Gilg tomg@cv.hp.com
nazgul@alphalpha.com (Kee Hinckley) (03/13/91)
In article <9103112113.AA20443@hpcvxph.cv.hp.com> tomg@HPCVXPH.CV.HP.COM (Thomas J. Gilg) writes: >pskq2_91 is a formal Domain/OS release available from Apollo to be installed >on top of SR10.3 Contact your Apollo sales rep I suppose. Offhand, what are the chances it would run on 10.2? -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
sl11@prism.gatech.EDU (LIEBESKIND,SUSAN H) (03/14/91)
Several weeks ago, Brian Hackley of Apollo sent me the magic number with
which you can order the media for the SR 10.3 PSK for X11R4
LFZ00BAD
The customer service person I spoke with tried to tell me that the PSK was
NOT out, and would NOT be out for some time. But, when I gave the fellow on
the phone this order number to look up, all the needed info appeared. Be
persistent if the representative you speak with tells you the same story.
The kit IS available now.
Thanks again to Brian for that information.
Susan Liebeskind
--
LIEBESKIND,SUSAN H
Georgia Tech Research Institute, GTRI/CSITL, Atlanta Georgia, 30332
uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!sl11
Internet: sl11@prism.gatech.eduHannu.Martikka@lut.fi (Hannu Martikka) (03/18/91)
>>>>> On 12 Mar 91 21:48:19 GMT, nazgul@alphalpha.com (Kee Hinckley) said: Kee> In article <9103112113.AA20443@hpcvxph.cv.hp.com> Kee> tomg@HPCVXPH.CV.HP.COM (Thomas J. Gilg) writes: >pskq2_91 is a formal Domain/OS release available from Apollo to be installed >on top of SR10.3 Contact your Apollo sales rep I suppose. Kee> Offhand, what are the chances it would run on 10.2? Can I get X11R4 (Hp/Apollo release not mit) to 10.2? We do have 10.3 (tapes) but none have ever looked at them, since we have some programs that might have problems in 10.3, so can I just extract X11R4 from 10.3 and install it somehow. Does anyone know if there are any major diffs between mit X11R4 and Hp/Apollo X11R4 release? (speed,etc...) -- Regards from Goodi ______________________________________________________________________________ Internet: Hannu.Martikka@lut.fi / \ Bitnet : GOODGULF@FINFILES // \\ \-\-\-\-\-\-\ oh5lhh Hannu Martikka, Punkkerikatu /// \\\ | on 70cm 7C 36, 53850 Lappeenranta,SF /// | \\\______________|_________________________ Tel. 953-251446 | :) Lappeenranta University Of Technology | ------------------------------------------------------------------------------
nazgul@alphalpha.com (Kee Hinckley) (03/19/91)
In article <HANNU.MARTIKKA.91Mar18160820@kannel.lut.fi> Hannu.Martikka@lut.fi (Hannu Martikka) writes: >Can I get X11R4 (Hp/Apollo release not mit) to 10.2? Apollo has had an SR10.2, borrow-mode R4 server for over a year. However as far as I know they've never released it, in fact I don't believe it was even created as an official product. I don't know whether or not someone at Apollo could be talked into releasing it unsupported. >We do have 10.3 (tapes) but none have ever looked at them, since we >have some programs that might have problems in 10.3, so can I just >extract X11R4 from 10.3 and install it somehow. The 10.3 tapes don't have the R4 server. It's a seperate PSK which only runs on 10.3. >Does anyone know if there are any major diffs between mit X11R4 and >Hp/Apollo X11R4 release? (speed,etc...) The Apollo version should be faster since it uses display specific drivers. How much faster probably depends on the type of hardware. -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
rn@ap.co.umist.ac.uk (bob nutter) (03/25/91)
In article <1991Mar19.031439.8944@alphalpha.com>, nazgul@alphalpha.com (Kee Hinckley) writes: |> >Does anyone know if there are any major diffs between mit X11R4 and |> >Hp/Apollo X11R4 release? (speed,etc...) |> The Apollo version should be faster since it uses display specific drivers. |> How much faster probably depends on the type of hardware. In our experience, using MIT X11R4 on mono dn3/4k's at sr10.1, the mfb driver is quicker than the gpr one, or at least it seems to interact better for us. Whether this applies to Apollo X11R4 as well I don't know. bob ------------------------------------------------------------------------------- bob nutter, computer officer | UMIST dept of computation | "Be Excellent To Each Other!" po box 88 manchester m60 1qd uk | -Bill & Ted tel:+44 61 200 3386 | email:b.nutter@umist.ac.uk |