whm@arizona.edu (Bill Mitchell) (08/05/87)
I posted a note a couple of weeks about how when running xterm on a VAX
using a 3.2 Sun as display server, everything would occasionally die and
I'd end up back at the Sun's shell prompt. There was some speculation
that this is due to not mousing enough and indeed, I often go for a long
period of time without mousing. But, I've often gone for an hour or so
of typing with no mousing and no problems and then turn around and have
X die three times in a hour (pretty frustrating when you've got a few
windows up, are logged in on several machines, and have an Emacs or two
cooking).
I run xinit with no arguments, putting up a local xterm window and then using
one or more rsh xterms as required. When it goes belly-up, the rsh xterms
disappear, taking their window them, the local xterm disappears in a similar
fashion, and then the shell prompt appears.
So anyway, I'm still looking for a solution (and still a little hesitant to
start pushing X locally), so if you've got any new ideas, I'm all ears, even for
things like putting three rings of colored sand around my Sun.
Bill Mitchell
whm@arizona.edu
{allegra,cmcl2,ihnp4,noao}!arizona!whmwhm@arizona.edu (Bill Mitchell) (08/06/87)
In response to my earlier query, someone had mentioned that maybe I was getting auto-logged out, but I checked our VAX shell and there's no autologout in it. jkh@violet.berkeley.edu suggested this in response to my second query and this time I thought to check my Sun xterm shell and sure enough, there was an autologout=60 in there. Setting autologout=1 would make things disappear just like usual, so silly me, autologout looks like the culprit. It appears if you've got a program running in the foreground, you aren't auto-loggedout. Thus, rsh vax xterm ... sticks around all day, but if you rsh -n vax xterm ... & the autologout gets you.
bob%aargh.cis.ohio-state.edu@osu-eddie.UUCP (Bob Sutterfield) (08/06/87)
In article <1856@megaron.arizona.edu> whm@arizona.UUCP (Bill Mitchell) writes: >... when running xterm on a VAX using a 3.2 Sun as display server, >everything would occasionally die and I'd end up back at the Sun's >shell prompt. There was some speculation that this is due to not >mousing enough ... So anyway, I'm still looking for a solution..., >so if you've got any new ideas, I'm all ears, even for things like >putting three rings of colored sand around my Sun. Well, in Ohio we don't use colored sand so much, so we had to find other possibilities :-). The not-mousing-enough speculation was mine, but was inaccurate. I posted <3802@osu-eddie.uucp> on 14 Jul describing our results, but perhaps it didn't make it to you. The cause in our case was a locally popular extended csh that had an autologout option that was set "on" by default. If a cshe has no activity for an hour then it will exit. For a user on a dumb terminal this was fine, because it meant he had walked away and forgotten to log out, and it solved part of the `untamo'-type problem. However, in a window system where some windows lie fallow for long periods of time (specifically the console xterm) it wasn't so helpful. So our solution is for users of cshe to "unset autologout", at least in the console xterm. There are a lot fewer frustrated X users around here now. I don't know if this will affect you because the csh extensions are local and you may not have something similar. But at least in our case, the cause was not that the user wasn't mousing enough, so you might want to look elsewhere. Perhaps colored sand isn't a bad idea - but be careful of the cooling fan. -=- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@ohio-state.{arpa,csnet} or ...!cbosgd!osu-eddie!bob soon: bob@aargh.cis.ohio-state.edu
wesommer@ICARUS.mit.edu (William E. Sommerfeld) (08/07/87)
In article <3911@osu-eddie.UUCP> bob@ohio-state.ARPA (Bob Sutterfield) writes: >The cause in our case was a locally popular extended csh that had an >autologout option that was set "on" by default. If a cshe has no >activity for an hour then it will exit. For a user on a dumb terminal >this was fine, because it meant he had walked away and forgotten to >log out, and it solved part of the `untamo'-type problem. > >However, in a window system where some windows lie fallow for long >periods of time (specifically the console xterm) it wasn't so helpful. >So our solution is for users of cshe to "unset autologout", at least >in the console xterm. There are a lot fewer frustrated X users around >here now. Sounds familiar. For a while, most of the terminals here at Athena were vt100/vt220 or equivalent, wired up to timesharing 11/750's. Someone put in the autologout option into /bin/csh (along with a bunch of other options, such as filename completion, etc.) for their benefit. Those who regularly used X (in general, those who knew what they were doing) generally tripped over the autologout problem pretty early, realized what had happened (or had it explained to them), and fixed it by adding an "unset autologout" to their .login file. As workstations started to become more common, their use by naive users increased, and there were a lot of bug reports of the form "The system just logged me out without warning -- all my windows went away". It was then decided to remove this option. The real problem was that it took a long time for the new /bin/csh to get propagated to all the systems. - Bill