Grex Helpers Conference

Item 84: Grex System Problems Item

Entered by i on Fri Sep 24 02:44:33 1999:

268 new of 292 responses total.


#25 of 292 by mooncat on Thu Sep 30 17:39:47 1999:

i don't know if this is connected to Merit or not, however, U of M had
a fire (apparently) in the computer area, so they've been trying to deal
with this.  I can't get into my U of M mail (apparently only a few boxes
were destroyed and thus some people can get access and some can't.). That
may have a connection to the Merit problem. <shrugs> dunno.



#26 of 292 by rcurl on Thu Sep 30 17:52:43 1999:

That's probably it. 


#27 of 292 by mcnally on Thu Sep 30 18:19:07 1999:

  Yep..  According to reports, a bank of batteries in the Computing Center
  building caught fire and began leaking acid.  Fire crews shut down power
  to the building and it's taken a while to straighten things out..


#28 of 292 by flem on Thu Sep 30 20:09:48 1999:

That would explain it.  I can't get to my email either. 


#29 of 292 by otaking on Thu Sep 30 20:56:36 1999:

When I came into work this morning at UM North Campus, I was told that some
idiot severed a major internet connection with a backhoe in Ohio. The
university lost e-mail and internet access for the entire morning because of
it. I could only access UM sites from here until noon.


#30 of 292 by mcnally on Thu Sep 30 22:54:57 1999:

  That also happened.  (Backhoes:  nature's most fearsome predator of the
  helpless fiber-optic cable..)  If you're interested in reading about that
  one there was an article in today's (9/30) Wired News..


#31 of 292 by arvindm on Thu Sep 30 23:19:15 1999:

pass


#32 of 292 by keesan on Fri Oct 1 01:37:18 1999:

Dpfitzen reports a problem with getting disconnected when trying to send
e-mails.  Has this happened to anyone else (using Pine)?  I suggested that
it might be caused by hitting Alt-X (a Procomm command to hang up) instead
of Ctl-X, but would like to know if other people are getting disconnected
while sending e-mail with Pine.


#33 of 292 by aruba on Fri Oct 1 05:23:19 1999:

OK, this is really weird.  I tried to send mail just now and found that
somehow my .pinerc had changed, with a line added that said
  default-fcc=nagraj@diversion.com
THis caused pine to reject my mail, because I have no save folder called
"nagraj@diversion.com".  My .pinerc is permitted 644, so no one else should
have been able to touch it, and I don't remember ever seeing that address
before (and it isn't in my current mail file or any of my saved mail files).

I think I'll change my password, though as far as can tell from "last", no
one else has logged in as me today.  Any idea what happened?


#34 of 292 by pfv on Fri Oct 1 05:27:39 1999:

see 21
see 22

Something funky has been happening that "isn't a problem".

Additionally, I'm experiencing periodic "seizure", which I can't explain
where: everything runs fine, then exverything seems to halt, no text, no
echo, no answer to AYT and even pinging grex tends to fail.

Seems likely to be two distinct problems, but oops - prolly not "problems"


#35 of 292 by tsty on Fri Oct 1 07:06:39 1999:

fwiw   
[No name] (DIVER-HST)

   Hostname: DIVERSION.COM
   Address: 207.126.100.96
   System: ? running ?

   Record last updated on 18-Oct-96.
   Database last updated on 30-Sep-99 04:34:27 EDT.



#36 of 292 by gull on Fri Oct 1 15:04:46 1999:

ZDNet article on the fiber cut:
http://www.zdnet.com/intweek/stories/news/0,4164,2343896,00.html

Article posted to mtu.resnet, forwarded from Jeff Ogden, describing in
detail the UM fire: http://www.tech.mtu.edu/~dmbrodbe/fire.txt


#37 of 292 by keesan on Fri Oct 1 19:16:35 1999:

I just got disconnected while in the Pine index and am now getting occasional
garbage in the middle of blank lines
Help (for more help), pine (fm
                              i   ia
                                   r 4 n

The above is typical.


#38 of 292 by mdw on Sat Oct 2 03:02:37 1999:

I got to look at CCB several times -- once just after the fire, when
there still a very thick haze of smoke inside (we went in, turned power
off on the IFS machines, and got out of there), and again late today, to
salvage non-critical hardware like power cables and the like.  The smoke
was almost entirely gone today when I looked, so I was able to wander
around back to the UPS area and look at where the fire started.  It took
me a bit to find it -- I had never been back there before, and there is
a maze of low down concrete rooms which house the boilers and other bits
of ventilation system.  I knew I had finally found the room, though,
when I found one which had a *black* ceiling.  There was a lot of broken
glass in there, as apparently all the ceiling fixtures had exploded in
the heat (wonder where all the mercury went?)

The UPS system has a large number of big blue 70's style black boxes,
with various sorts of cryptic labels on them.  I never did figure out
which ones (if any) housed the diesels, but along the back wall, there
were a number of big anonymous looking boxes that turned out to be the
battery racks.  I could tell because the furthest one back was *the* one
that burned.  The blue paint had turned into white ash, the shiny metal
under it was black with corrosion and soot, and of course, the batteries
were still inside.  The ones in the bottom half of the case didn't look
too bad, they were still white plastic cases on shelves.  Along the top
half, though, the cases were severely melted and carbonized, and had
mostly slumped into a thick lumpy glaze on the tops of the batteries.
(The shelf probably kept the bottoms from getting so hot).  The wiring
conduits above these batteries had definitely undergone an unusual
experience.  These were basically aluminum tubing with really heavy
gauge wire rope conductors inside.  Presuambly they had once been
insulated, but evidently where the insulation had melted, the conductors
had touched the tubing, and vaporized it.  So the tubing basically had
very elongaged and slightly irregular holes chewed along one side of it.

There were several other UPS boxes next to the battery case, and while
these weren't in nearly as bad a shape as the battery case, they showed
distinct signs of damage.  These cases happened to have small control
panels on them, with some sort of meter (voltage?  current?) consisting
of a glass faceplate and black plastic framing it.  Much of this had
melted and deformed in the heat, giving it a very surrealistic
apparance.

Apparently, the fire put itself out.  Lucky us.


#39 of 292 by other on Sat Oct 2 04:36:48 1999:

i'll bet that the sheer volume of smoke and already oxidized gases choked out
the fire by starving it of sufficient oxygen to keep it going.  between that
and the soot and acid gas, it makes a good argument for keeping these things
in a *poorly* ventilated, cool basement room.


#40 of 292 by aruba on Sat Oct 2 13:28:40 1999:

Thanks for the description, Marcus.


#41 of 292 by tod on Sun Oct 3 14:34:03 1999:

re #38
Didn't the data center have a form of automated fire protection?
I understand that housing such a vast quantity of batteries in one
place constitutes a severe fire hazard.
..


#42 of 292 by davel on Sun Oct 3 15:49:27 1999:

Yesterday, repeated efforts to log in connected to the termserver (got the
"Welcome to Grex!  (it may take a few seconds to connect)" message), followed
quickly by a disconnect.  An attempt to come in over the net got "connection
refused".


#43 of 292 by orinoco on Sun Oct 3 22:38:51 1999:

I had an interesting error when I was on Grex this morning.  I was typing a
response somewhere, when suddenly every keystroke just made the computer beep
at me.  I have _no_ idea if this was the fault of the computer I was on, or
of Grex, sinec I had to run anyway and I'm now logging on from a different
computer.  Who knows...


#44 of 292 by scott on Sun Oct 3 22:42:16 1999:

Probably your own computer, complaining about something.  Was it Windows? 
Had you hit the ALT key by accident?


#45 of 292 by gelinas on Mon Oct 4 00:32:59 1999:

No, the data center did not have a fire suppression system.  Or so I've
heard; I've not actually investigated it myself.  The new data center
does have a sprinkler system.


#46 of 292 by mdw on Mon Oct 4 01:37:50 1999:

Actually, what the old building needed was a smoke alarm.  I think it
did have sprinklers in some places, but not in the UPS bunker.  As it
happened, the fire *did* announce itself to public safety -- when the
temperature and power went out of bounds, and unfortunately after a lot
of smoke had been generated and a lot of wiring melted.  The nice thing
about the old building was that the UPS was in fact in a concrete bunker
well separated from the building, and concrete does not burn very well
at all.  It's too bad it shared the same ventilation system, otherwise,
the disaster could have been a lot less annoying.

I doubt any standard sort of fire suppression system would have done
much good for the UPS.  Sprinklers would have just dumped water on the
batteries, which probably would have then shorted out that much faster
and more violently.  A halon suppression system (which is now illegal to
install) probably wouldn't have made any difference to the short, but
would have interacted in a very interesting but probably much more
deadly fashion with the subsequent electrical fire (it probably would
have decomposed, releasing much greater amounts of chlorine, flourine,
and etc., which would have greatly hampered recovery efforts and
probably damaged more equipment.)


#47 of 292 by gull on Mon Oct 4 13:01:54 1999:

Really, until the short condition (assuming that's what it was) was removed,
nothing was likely to put out the fire.  Were the batteries equipped with
fusable links, I wonder?


#48 of 292 by rcurl on Mon Oct 4 15:15:20 1999:

I've  heard of exploding lead-acid batteries in cars, and have thought
it was a hydrogen-air explosion, but come to think of it, the plates
in the battery are quite close together, and a internal short is quite'
possible. That should lead to some fireworks in a fully charged battery.
There would be no possible external protection against such a failure,
except to not have the explosion of one battery short or disrupt others. 
I would be interested in what was the initial failure mode, if it can be
determined, as I use lead-acid batteries in more than my car.


#49 of 292 by gull on Mon Oct 4 19:08:49 1999:

It seems like a hydrogen/air explosion in any one battery wouldn't produce
enough heat for long enough to start a fire, but I could be wrong.  The
continual smokey smoldering that it sounds like occurred sounds more like an
electrical fire to me.  We had a maintenance-free lead-acid battery explode
on a travel trailer once, and there was no sign of burning or melting
damage.  (The battery was outside, on the A-frame, next to the propane
tanks.)


#50 of 292 by aruba on Tue Oct 5 01:39:21 1999:

I had trouble dialing in on 761-3000 tonight; modem connected but I never got
the message from the terminal server.  761-5159 worked fine.


#51 of 292 by mdw on Tue Oct 5 02:11:24 1999:

I haven't heard any definitive theory as to the cause yet.  One of the
theories bandied about was that "something got dropped into the
batteries".  I looked at the actual battery case.  It's totally
enclosed, but the top plate of the box is perforated with numerous small
holes.  Inside the cabinet, the battery posts (lead, presumably) stick
up, and are connected to each other with some sort of grey silver metal
braid.  I suppose a wire dropped in *just right* might slip through the
holes, and could in theory land across 2 terminals and short out, but I
kind of doubt it.  For one thing, I think a pieces of wire small enough
to slip through the holes would probably have been small enough to
vapourize itself without much harm to anything else.  For another thing,
it would have had to drop through the holes vertically, then somehow
twist itself around horizontally *and* manage to make a really good
connection with 2 different bits of metal.  The battery terminals would
have been pretty well oxidized, and they are, after all, pretty small
targets in an area consisting mostly of plastic and painted metal.

I don't believe there were any fusible links.  Fusible links are
basically just low tech fuses, being nothing more than a short length of
under-gauge wire; if the curcuit overloads, the theory is, the fusible
link is the bit of wire that will go first, and it's generally located
in an area where it's easier to fix, and less likely to cause damage.
Fusible links are not suitable for high voltage or high current
applications.  For instance in a car the primary circuit between the
battery & the starter generally has no fuses or circuit breaker of any
sort.  The current draw is too high to make a fusible link practical,
and in fact the starter has a low enough resistance that it's not really
feasible to tell the difference between starting the car, and a dead
short in the solenoid.

For a stationary installation like the computing center, either fuses or
circuit breakers would be used.  Of the two, circuit breakers are much
more likely.  For power loads of this magnitude, blowing either a fuse
or a circuit breaker would have been a very dramatic affair.  It would
have generated a loud noise, lots of ozone, UV radiation, and vapourized
stuff.  (oil, insulation, wahtever.) Those big fancy power panels aren't
just for show; one of their primary purposes is to protect the operators
from the potential fireworks if a breaker under full load should go.  In
a big electrical substation, the breakers are kept outdoors along with
the transformers in an area fenced off from ordinary mortals.  The
computing center wasn't big enough to rate the outdoors display; but it
did have many large breakers and other controls located along a wall.
These would probably have done a great job of protecting the rest of the
building if a short had happened almost anywhere "downstream" of the
UPS.  I didn't see any evidence that the battery cases themselves had
any sort of protection - and I'm not sure it would have been any more
practical there than in a car in the starter circuit.


#52 of 292 by n8nxf on Tue Oct 5 11:38:16 1999:

The one exploded car battery I saw seemed to have done so because of an
internal spark caused by plates loosening from their buss bars.  In the
quest to save lead they had used less in these junctures and they can
break and spark when there is a large current draw.


#53 of 292 by mjb on Wed Oct 6 19:50:35 1999:

Back in Junior year of High School, I was in Auto Shop class, and the 
idiot they hired to be 'teacher's aide' hooked up a car battry to a charger,
backwards.  The explosion that followed only a few minutes later was quite
violent.  The case had cracked all around the battery, and the core was
visible where the two halves of the case had separated.  Acid flew every-
where, but, fortunately, noone was anywhere near the area when the explosion
occurred.  What I remember most was how loud that explosion was...Batteries
store a great deal of energy, and can be quite nasty......


#54 of 292 by n8nxf on Fri Oct 8 11:21:11 1999:

The explosion was caused by hydrogen and oxygen recombining into water, not
by energy stored in the battery.  Being connected backwards it split a lot
more water than it usually would as current flow through the charger was
probably way in excess of what it was rated for.


#55 of 292 by krj on Fri Oct 8 20:09:47 1999:

Back to Grex problems.  Twice today I have attempted to mail a 26K 
file to two Grexers and the mail has failed with the following 
uninformative message:
 
   ----- Transcript of session follows -----
... while talking to grex.cyberspace.org.:
>>> DATA
<<< 554 SMTP-MAIL: died on signal 11
554 mooncat@cyberspace.org,gypsi@cyberspace.org... Service unavailable
 
Have I discovered an intentional limit on e-mail size, or is this a bug?


#56 of 292 by rcurl on Fri Oct 8 21:52:01 1999:

I don't think #54 is correct. A battery generates hydrogen and oxygen by
electroysis of the battery acid during charging, after the battery is
fully charged. It does not generate the gases on discharge, which is
what would be happening with the backward charger. I would expect that
the explosion would result from heat generation.


#57 of 292 by n8nxf on Sun Oct 10 00:38:58 1999:

I think you are correct.  However, the space between the plates and the top
of the case is filled with oxygen and hydrogen gas unless the caps are
removed.  The exploded batteries I've seen had the cases ripped apart above
the electrolyte level but were intact below this level.  Had the energy
for the explosion come from the stored charge, I'd expect the lower part
of the case to be shredded.  I have also noticed gas bubbles while
charging discharged lead-acid batteries before they were fully charged.


#58 of 292 by krj on Sun Oct 10 04:26:22 1999:

Could I please get a comment on my resp:55 from a staff member?


#59 of 292 by pfv on Sun Oct 10 04:39:20 1999:

        They're all too busy debating fires and batteries.


#60 of 292 by rcurl on Sun Oct 10 05:31:39 1999:

The explosion can be a result of a mix of things. The only gases generated
in the battery is a stoichiometric mixture of H2 and O2, which really
goes bang. The shorting of the battery can generate enough heat to
probably melt connections and hence create sparks. Unless the battery is
ventilated, it will retain that explosive gas mixture for quite a long
time, from previous charging. And yes, the overvoltage used for charging
can indeed produce some of the gas mixture even before the battery is
fully charged. 


#61 of 292 by drewmike on Sun Oct 10 18:39:27 1999:

Item 65.


#62 of 292 by aruba on Sun Oct 10 19:29:16 1999:

Ken, I believe the limit on e-mail size is 100K, so I don't know why your mail
was rejected.


#63 of 292 by richard on Tue Oct 12 15:02:14 1999:

is there any way to expand the time (30 seconds) grex allows for one to
login once one gets a login prompt-- when there is a long que and one
leaves the grex window open while the que is running, 30 seconds is
often not enough time to catch the login while it is up.  Maybe 60
or 90 seconds would be better.


#64 of 292 by mcnally on Tue Oct 12 16:20:12 1999:

  How about if I increase it to 120 seconds, retroactive for the past
  year or two? 

  <poof!>  It's done.



  (and all without having root, too!)


#65 of 292 by danr on Tue Oct 12 16:40:16 1999:

Impressive.


#66 of 292 by richard on Wed Oct 13 16:31:36 1999:

oh, is it 120, well it seems like 30 seconds, maybe it should be
upped to 180 seconds or 240.


#67 of 292 by rcurl on Wed Oct 13 17:07:37 1999:

Maybe you should pay attention. 120 seconds is two minutes when other
people in the queue are being forced to wait because you are not paying
attention. 


#68 of 292 by richard on Wed Oct 13 18:10:57 1999:

also, is there a way to turn off !talk requests without turning
of !tels?  Or at least limit the number of times you can get paged
for a !talk request to maybe twice.  I hate it when complete
strangers, usually new users, !talk me and I get paged for what seems
like forever.  


#69 of 292 by pfv on Wed Oct 13 18:22:49 1999:

        I thought this was a "non-problem" aka: "non-issue"..?


#70 of 292 by remmers on Sat Oct 16 23:39:00 1999:

I think 120 seconds is reasonable.  It *does* beep when it gives you the
login prompt; this is supposed to get your attention.

As for allowing tels and disallowing talk requests - I'm not sure if
it's possible currently, but with the architecture of the write/tel
program I don't think it would be too difficult to implement.
I think that it's possible to get part of the effect you want by doing
"mesg ne" and then specifying in your .yeswrite file (if that's the
name) a set of people who are allowed to write/tel you.


#71 of 292 by eprom on Sun Oct 17 06:57:56 1999:

how bout adding a couple more beeps 
If im in another window and I have a window to m-net open and idling
it will beep three time....it tends to get my attention better than a
single beep...


#72 of 292 by goose on Mon Oct 18 10:03:12 1999:

Please no more beeps.


#73 of 292 by mcnally on Mon Oct 18 23:00:52 1999:

  Dialed in this evening to 761-3000 (no idea what line I wound up on)
  but got the "connecting.. this may take a minute" message followed
  by NO CARRIER.  Happened several times running -- perhaps one (or more)
  of the dial-ins is messed up?


#74 of 292 by scott on Mon Oct 18 23:28:33 1999:

Nope, inetd had died somehow.


#75 of 292 by tpryan on Tue Oct 19 01:59:10 1999:

        I had the same dial-in problem earlier this evening.


#76 of 292 by mcnally on Tue Oct 19 04:40:43 1999:

  Hmmm..  I'd've thought that would've prevented me from ssh'ing in,
  too -- or was it fixed by the time I posted #73?


#77 of 292 by gull on Tue Oct 19 05:05:52 1999:

sshd generally isn't run from inetd.  The reason is that inetd only launches
a program when there's an incoming connection.  That's a bad idea for sshd,
because it needs to generate an encryption key when it's started, and that
can take 15 seconds or so.  So usually it's just allowed to run all the
time, independantly of inetd.


#78 of 292 by mcnally on Tue Oct 19 05:27:57 1999:

  Makes sense..


#79 of 292 by scg on Tue Oct 19 05:42:05 1999:

Running sshd and inetd independantly is also useful if one of them dies, since
it's then possible to access the system using the other one.


#80 of 292 by jazz on Tue Oct 19 15:27:17 1999:

        Naah, that's what modems're for! :)


#81 of 292 by richard on Tue Oct 19 15:28:48 1999:

I dont really want to set up .yeswrite or .nowrite files because I dont
mind getting tels, and it feels anitsocial in this environment to be
using such files.  I just want to permanently block any !talk requests
because anybody !talking me should really be !teling me, and I have
rarely ever had somebody !talking me from offsite.  there isnt a way
to do that?


#82 of 292 by pfv on Tue Oct 19 15:57:19 1999:

        Damn, that's racist and antisocial - you should let EVERYONE tel &
        talk to you! At *ALL* times!

        Sheesh..


#83 of 292 by richard on Tue Oct 19 16:28:40 1999:

its not racist, I dont mind if an Indian user !tels me-- some of them
are nice-- I just dont want to turn off !tels just because I want to
turn of !talk


#84 of 292 by pfv on Tue Oct 19 17:12:06 1999:

        Separationist!


#85 of 292 by omni on Tue Oct 19 17:30:16 1999:

  I'm getting sick of being constantly !tel'ed and talked. I come to grex to
do chat and I damn well aint going to start now. Wanna talk to me, send me
an e-mail


#86 of 292 by remmers on Tue Oct 19 17:35:23 1999:

I think being able to turn off talk requests without turning off tel's
would be reasonable.  I'd email janc about it - he's the
author/maintainer of the write/tel stuff.


#87 of 292 by gull on Tue Oct 19 22:26:05 1999:

At very least, someone should hack the 'talk' source so it only pages you
*once*, instead of repeatedly.  It's a very rude program.


#88 of 292 by pfv on Tue Oct 19 22:33:17 1999:

        It's garb- errr.. "it's not a problem" .."it's not an issue"


#89 of 292 by drewmike on Wed Oct 20 02:35:59 1999:

(Hey, did anyone else notice that Peefv never mentioned any specific
ethnicity, which Richard instantly pegged as Indian? Hell's yeah, I still got
the instigator in me...)


#90 of 292 by mooncat on Wed Oct 20 12:41:19 1999:

It may not be a probelm or an issue, but I would like to be able to turn
 !talk off completely... I never use the program, and I don't like it.



#91 of 292 by pfv on Wed Oct 20 14:51:18 1999:

        Gee, youse guys make it sound like it's your RIGHT to control 
        the contents of yer CRT..

        (while you are modding, set it up like bbs or party: user defined
        filter-insertion)


#92 of 292 by gelinas on Thu Oct 21 01:17:12 1999:

Edit your .login file and make sure it has the lines

        setenv MESG mesg
        mesg n

If you change your mind, change that final character to a "y".


#93 of 292 by davel on Thu Oct 21 01:47:24 1999:

... or just do "mesg y" on the fly to change it for one session.


#94 of 292 by gelinas on Thu Oct 21 03:22:31 1999:

I wish I could backspace at the login: and password: prompts.  And I'm
not all that bad a typist.


#95 of 292 by gull on Thu Oct 21 04:33:13 1999:

You have a problem with your terminal settings.  Your backspace key is
probably set to send ^?, while login expects ^H.


#96 of 292 by gelinas on Thu Oct 21 04:49:33 1999:

Nope.  I can solve *that* problem.  But when I do, grex reports, "Not allowed
to modify login" or words to that effect.  And my terminal sends <CTRL>H,
which gets displayed as "^H", of course.


#97 of 292 by gelinas on Thu Oct 21 04:57:22 1999:

Here 'tis:

} (ttyr6) grex login: gelasdfk^?^H^?^?^?^H^H^H^H^H^H^?^?^?^?^?^?
} You may not change $L0
} gelas's Password:

Hmmm... This gives me an idea.


#98 of 292 by gelinas on Thu Oct 21 05:01:11 1999:

That did it:

} login: abcd^H^H^H^Hgelinas
} gelinas's Password:

So I just have to carefully count how many <CTRL>Hs I need.


#99 of 292 by mdw on Thu Oct 21 10:32:22 1999:

If you type whitespace on the login line, the extra parameters get
converted to environment variables L0, L1, etc.  Login checks for
various extraneous environment variables (such as LD_LIBRARY_PATH) that
have security implications, and rejects them.  It uses the pattern L*,
which has the side-effect of rejecting L0,L1 as well.  It might not be
obvious, but L0, L1 do have security implications -- sometimes people
who fat finger their password manage to get it into the login prompt,
and we don't want to export passwords or parts thereof into the
environment.


#100 of 292 by jep on Thu Oct 21 13:46:58 1999:

I use ^U if I make a mistake on the login: or password: line, then I 
re-type it.  It works great.


#101 of 292 by scott on Thu Oct 21 13:59:26 1999:

(Control-U is "erase line" in UNIX)


#102 of 292 by janc on Thu Oct 21 15:39:58 1999:

I don't understand why login can't treat *both* ^H and ^? as delete
characters when logins and passwords are being entered.  That seems the
only sane solution.  Only catch is that the change over would annoy
anyone whose password includes a ^?, but any such person pretty much
deserves it.


#103 of 292 by janc on Thu Oct 21 15:41:37 1999:

Oh, and I agree that there should be a way to turn off talk requests
seperately from write/tel requests.  Someone who has a couple days free
should write this sometime.


#104 of 292 by mooncat on Thu Oct 21 16:22:28 1999:

Back there for the person who suggested 'mesg n' for the .login file...
Doesn't that turn off *all* messages?  I don't want that.  I like
tels, and I tend to talk to many people at once via tels while
in party or something.  What I don't like, or use is !talk... 



#105 of 292 by other on Thu Oct 21 19:39:59 1999:

if i remember correctly, the manual for mesg or for nowrite/yeswrite includes
a flag to select between tel/talk/chat/all requests.  i have a yeswrite file
and am flagged for tels only.


#106 of 292 by other on Thu Oct 21 19:40:38 1999:

          mesg ne -p t -r y -b y -m l


#107 of 292 by jazz on Thu Oct 21 22:43:32 1999:

        mesg -2 -m -a -n -y -o -p -t -i -o -n -s.


#108 of 292 by other on Thu Oct 21 23:33:41 1999:

i dunno, i spent half an hour and got it to do exactly what i wanted...


#109 of 292 by orinoco on Fri Oct 22 00:13:22 1999:

For those who were asking somwhere about making telk requests not beep at you,
the manual makes it look like   mesg -b y   will do that.


#110 of 292 by pfv on Fri Oct 22 01:04:46 1999:

        I think what they said was they didn't want to: see, hear,
        smell, taste, touch or SENSE "talk requests".

        OTOH, they wanted to live without losing telegrams.


#111 of 292 by krj on Fri Oct 22 06:18:14 1999:

Question from arabella: has anything changed which would break the 
combination of "setenv TERM ansi" and the more pager?
Today Leslie starts getting the message: "WARNING: terminal cannot 
scroll backwards".  We don't think she changed anything in her 
configuration files.


#112 of 292 by jor on Fri Oct 22 14:12:40 1999:

        it's not April Fool's??

 10:11am  up 10:26,  1 user,  load average: 0.08, 0.01, 0.00
User     tty       login@  idle   JCPU   PCPU  what
jor      ttys0    10:06am            5         w 


#113 of 292 by goose on Fri Oct 22 15:16:12 1999:

Yeah, something's not right.  First a traceroute stopped at blackrose.org,
then a couple minutes later it was hung up somewhere within cyberspace.org.


#114 of 292 by krj on Fri Oct 22 16:20:36 1999:

Late last night I found that I could not telnet to Grex from 
Michnet or tir.com; but when I dialed in I found lots of other people
telnetted in.  Today I can't telnet in from Michnet, tir.com, or msen.com,
and apparently no one else can telnet in either.


#115 of 292 by keesan on Sat Oct 23 00:42:54 1999:

Aruba says the two of us are alone on grex now.  So why is it taking a minute
or two to send off a short e-mail?  All evening.


#116 of 292 by kaplan on Sat Oct 23 02:29:46 1999:

Grex's connection to the Internet is down.  Blame Ameritech.  Why is it taking
you minutes?  I don't know.  Do you mean it took a long time for your mail
client to accept the message?  Are you using pine?  If it's being delivered
to someone else on grex, that should be no problem.  But if you're trying to
deliver to someplace out on the Internet, it won't get there until the link
is fixed!


#117 of 292 by keesan on Sat Oct 23 03:21:16 1999:

Sending to someone outside of grex, took over a minute to go from 0% to 100%
using pine.  Where did it go?  Have not tried lynx recently.  There are four
of us dialled in at the moment, including someone in Party, signed on for 20
minutes.  Who are they partying with?  It takes two to party.


#118 of 292 by scg on Sat Oct 23 03:33:26 1999:

The long delay was probably something trying to do a DNS lookup, which
requires network connectivity.


#119 of 292 by aruba on Sat Oct 23 14:01:01 1999:

I had to wait several minutes at the "Welcome to Grex!  (it may take a few
seconds to connect)" prompt.  Does that have to do with the net connection
being down?


#120 of 292 by cmcgee on Sat Oct 23 16:24:07 1999:

This morning it took 2:45 (minutes and seconds) to get from the Welcome to
Grex to the login prompt using dial-in lines.  And all my mail that I sent
yesterday (which took over a minute per message to send) has mailerdeamon
messages this morning saying it didn't go through. 




#121 of 292 by krj on Sat Oct 23 17:25:29 1999:

Colleen, you're mail's not going anywhere until the Internet connection
is fixed.   There ought to be a large motd message explaining this, staff.
 
I confirm the two minute 45 second wait for a login prompt.


#122 of 292 by scg on Sat Oct 23 17:41:41 1999:

The delay waiting for the login prompt was DNS timing out.  That should be
fixed now.

Ameritech says they've found the ISDN problem, and are working on it.


#123 of 292 by scg on Sat Oct 23 20:51:28 1999:

Since it had been almost three hours since Ameritech had called to say 
it was a cable problem and was being actively worked on, I took a walk 
over to the Pumpkin  to see if I could find any evidence of that.

I started by walking around the building, looking for where the phone 
cables come out.  I had assumed they came out the front of the building 
and then went straight up Huron to the CO, but there don't appear to be 
any overhead phone cables in that part of Huron.

The phone lines come out of the side of the building, fairly near the 
back, and then head out behind the homeless shelter and the rental car 
place to Chapin Street.  From there, they head towards Huron.  I didn't 
have to follow them any farther than that.

Near the corner of Chapin and Huron, there were two Ameritech trucks, 
next to two utility poles, maybe 100 feet apart.  Running between them 
were a phone cable at normal height, and a another cable hanging down at 
about half the normal phone cable height.  On each pole was an Ameritech 
guy.  One of them was covered by a big cloth tent-type thing, so I 
couldn't see what he was doing.  The guy at the other end had both cables 
split out into the individual wire pairs, and appeared to be going 
through, pair by pair, and replacing one cable with the other.

It's currently drizzling and 43 degrees in Ann Arbor.  It didn't look 
like pleasant work.

Assuming they're doing what I think they were doing, we will probably 
lose each of the POTS lines briefly, one by one, as they work on this.  
I'm also guessing that we probably weren't the only customer affected, if 
they're replacing the whole cable.  I wonder if we lost any of the 
dial-in lines, without noticing.


#124 of 292 by senna on Sat Oct 23 23:11:38 1999:

The net is up, but everything is extremely slow.


#125 of 292 by ryan on Sat Oct 23 23:28:59 1999:

This response has been erased.



#126 of 292 by scott on Sun Oct 24 13:05:20 1999:

(Scott used to think ryan was cool)


#127 of 292 by jazz on Sun Oct 24 14:04:24 1999:

        Yeah, he should know sendmail is (or can be) load-sensitive.

        I was amazed when I ran the sources of most of the SMTP-forwarded mail
on m-net against the local user database.  It really was mostly from India
(or proxies like mason.ge.com that are American proxies for Indian sites).

        To be fair, most of the POP3 abuse on m-net *was* from within the
'States, though most of it wasn't in the local user address database either.


#128 of 292 by ryan on Sun Oct 24 20:09:51 1999:

This response has been erased.



#129 of 292 by keesan on Mon Oct 25 00:33:55 1999:

What does Fatal exception in module kernel.dll mean?  
I got this both times that I exited the graphics conference.
I am dialled in directly, using Procomm Plus/DOS.


#130 of 292 by jazz on Mon Oct 25 01:24:28 1999:

        It's a joke.

        UNIX doesn't call dynamically linked libraries "*.dll".


#131 of 292 by mcnally on Mon Oct 25 03:28:07 1999:

  more specifically, it's a dig at Microsoft Windows, which produces
  "Fatal Exception" errors with distressing frequency..


#132 of 292 by gelinas on Mon Oct 25 03:34:47 1999:

But both are neglecting to mention that the text is the 'exit message'
for the conference, much as diy (or is it micros?) says 

        Boop

when I leave it.


#133 of 292 by gelinas on Mon Oct 25 03:38:47 1999:

It's micros.


#134 of 292 by aruba on Mon Oct 25 13:24:30 1999:

Grex hung up on me at 9:15 this morning.  I was able to get right back on.


#135 of 292 by dpc on Mon Oct 25 14:03:50 1999:

Just a few minutes ago I was on a dialin and experienced the now
all too familiar disconnection without any reason.


#136 of 292 by richard on Mon Oct 25 15:25:00 1999:

does grex get a refund for the days or hours ameritech couldnt provide
the ISDN service it was paying for? would seem only fair



#137 of 292 by aruba on Mon Oct 25 17:28:13 1999:

Yes, I believe we are entitled to a refund.  Our ISDN lines cost $48.05 each
these days (including taxes and fees), so two days of service should be worth
(2/30)*($48.05) = $3.20.  Not sure if they'd pay for both days since we didn't
call them until Friday evening, and they had it fixed within 24 hours of then.

I'll give it a try.


#138 of 292 by jazz on Mon Oct 25 23:09:26 1999:

        
        Ameritech doesn't offer SLAs on those, do they?


#139 of 292 by keesan on Tue Oct 26 00:30:39 1999:

What is a dynamically linked library?


#140 of 292 by aruba on Tue Oct 26 01:20:50 1999:

Re #138: I don't know what an SLA is.


#141 of 292 by drew on Tue Oct 26 01:39:48 1999:

Re #139:
    It is a file containing commonly used functions particular to Windows.
Windows aps depend on these files being there rather than have the code
compiled into the main executable. In theory it cuts down on disk space useage
by eliminating the redundancy of storing the same routines once for each
application.


#142 of 292 by gelinas on Tue Oct 26 01:58:13 1999:

Further, you can update those functions by dropping in a new libary, a
new dll.

"SLA" is "service-level agreement", a statement of what you get for what you
pay.


#143 of 292 by albaugh on Tue Oct 26 04:44:13 1999:

SLA is most often discussed in terms of "support", up-time, response, 
24/7, the like.


#144 of 292 by aruba on Tue Oct 26 13:22:58 1999:

I don't know if we have a service-level agreement or not.


#145 of 292 by jazz on Tue Oct 26 19:11:12 1999:

        Right.  Some ISDN providers guarantee a certain amount of "up-time".
I don't think Ameritech does but it's certainly worth a shot - if Ameritech
did offer an SLA, it'd certainly be an entire month's credit GREX would be
getting insteadd of a few days'.


#146 of 292 by keesan on Fri Oct 29 17:39:52 1999:

Does UNIX also have GPFs?  Or only Windows?


#147 of 292 by scott on Fri Oct 29 18:23:23 1999:

Unix has "segmentation faults".  Same thing, a program tries to access
something that it isn't supposed to.


#148 of 292 by pfv on Fri Oct 29 20:13:38 1999:

        Well... hehe - Actually, unix/linux segfaults are probably
        easier to track down, since you can't ever be sure the GPF isn't
        directly due to the virus they call an OS..


#149 of 292 by mdw on Fri Oct 29 20:26:34 1999:

In unix, depending on the hardware, it's also possible to get an
"illegal instruction", or "bus" signal, or sometimes a "bad argument to
system call" or other interrupt.


#150 of 292 by pfv on Fri Oct 29 21:06:42 1999:

        Oy, geezus.. I had something like that "buss error" thing for
        awhile.. Damned catchall is as bad as a GPF.. Fixed it, but gad..
        It was depressing.


#151 of 292 by mdw on Fri Oct 29 22:27:15 1999:

What SIGBUS means is hardware dependent, but originally (on the pdp-11 &
vax), it meant an attempt to fetch a word on a non-word boundary -- an
"alignment" error.  Many modern machines can fetch words on non-aligned
boundaries (usually with a small performance penalty) so this is not as
likely.  The i86 can do unaligned fetches.  The 68000 and 68010 require
everything except bytes be at even addresses.  The 68020 and + can do
aligned data fetches, but still require that the stack & program counter
be even.  The sparc and power risc architectures can't do aligned
fetches.  Most powerpc implementations can do aligned data fetches (but
would require word aligned stack & instructions.)

In openbsd, apparently, you'd have to do a systemcall that somehow
invalidates the current stack or data segment; perhaps a call to mmap.


#152 of 292 by pfv on Fri Oct 29 23:13:42 1999:

        That was precisely the problem, too.. I had some code developed 
        (apparently) for DOS that relied on data being so & so arrayed..

        I was pretty sure it was that, but I was too tired that day to
        track it further.. Thankfully, another grexer took the reins and 
        slammed out a solution.. 

        Sorta' a Terrible-Day & then the Sun-Came-Out, thang..



#153 of 292 by gelinas on Sat Oct 30 00:17:33 1999:

(Marcus, did you sometimes use "aligned" above when you meant "unaligned"?)


#154 of 292 by mdw on Sat Oct 30 03:25:32 1999:

(Probably.  Are you studying to be a proof-reader?)


#155 of 292 by gelinas on Sat Oct 30 03:27:56 1999:

(No, it's a curse.)


#156 of 292 by hhsrat on Sat Oct 30 03:28:51 1999:

(Who's this General Protection dude, and why is everything always his 
fault? :)


#157 of 292 by tpryan on Sat Oct 30 15:45:31 1999:

        He took over from General Mills.


#158 of 292 by krj on Wed Nov 3 05:28:30 1999:

/a is almost full, someone in party was complaining about a lost 
participationf ile.


#159 of 292 by gelinas on Wed Nov 3 06:09:25 1999:

Well, it's at 100% right now, and my part file was merely corrupted.  I've
not deleted it yet.


#160 of 292 by aruba on Wed Nov 3 13:44:47 1999:

I lost my participation file to /a last night.


#161 of 292 by drew on Thu Nov 4 20:27:46 1999:

Time to put in one of those extra hard drives, and move half of /a to it.


#162 of 292 by richard on Fri Nov 5 20:42:22 1999:

Arrgh, are vandals attacking the /a drive again and filling it up?

Im getting a "bad participation file" message every time I log on, and
my conf bookmarks arent being kept upto date.  How can I fix?


#163 of 292 by other on Fri Nov 5 20:56:35 1999:

yeah, richard!  quit taking up so much drive space!     ;)


#164 of 292 by gull on Fri Nov 5 21:17:49 1999:

Keep a big junk file around, so when the drive fills up, you can delete it
and get enough space to save your participation file. ;>

(I'm kidding. Don't do that, it's not polite.)


#165 of 292 by goose on Sat Nov 6 21:34:05 1999:

HELLO STAFF?!?   /a ihas now become a chronic problem.......


#166 of 292 by krj on Sun Nov 7 06:46:19 1999:

Yes, the /a problem just trashed another participation file of mine.


#167 of 292 by sno on Sun Nov 7 15:29:40 1999:

Why can't a link be used for certain groups' home directory to a known
viable space on another volume.  It seems that waiting to repartition
doesn't fix the problem very fast.


#168 of 292 by jazz on Sun Nov 7 16:53:41 1999:

        I'd thought about that, or taking an additional step, and creating a
partition specifically for BBS partition files.  The only advantage here is
that the volume would be easier to manage when conferences are dismanted (if
they are here, I haven't seen it happen, but it was a mess on m-net) - it's
still a limited volume and will still fill up. :)


#169 of 292 by scg on Sun Nov 7 17:20:03 1999:

I switched newuser over yesterday so that it is now creating new accounts on
the /c partition instead.  I fell asleep in the middle of trying to clear up
space on /a.


#170 of 292 by jazz on Sun Nov 7 17:43:17 1999:

        It's a tough job, Steve.  Good luck.  Have you considered changing the
reaping times for inactive accounts?


#171 of 292 by hematite on Mon Nov 8 04:08:27 1999:

How does one get switch the new menu system back to the old one?


#172 of 292 by mcnally on Mon Nov 8 04:50:56 1999:

  Given that:

    1) the people who intentionally vandalize the system by filling up
       filesystems in a denial-of-service attack generally do so from
       new accounts, not long-established ones, and that
    2) "filesystem full" problems are of particular annoyance to conference
       participants..

  Would it be profitable to consider a home-directory allocation scheme
  where long-established accounts with a history of mostly stable disk-
  usage were grouped together on one or more partitions set aside for
  the purpose?  If the majority of long-time confer participants had home
  directories which were physically segregated away from newuser accounts
  it seems like many of the surprise filesystem problems which affect
  Picospan users could be diverted..




#173 of 292 by orinoco on Mon Nov 8 05:13:52 1999:

That sounds like it would just be inviting trouble, though.  Who would
determine who gets to be a "special user" with a stable directory?  Wouldn't
giving accounts which are likely to lose their participation files to new
users tend to make conferencing difficult for the people we're trying the
hardest to attract to the conferences?  

It seems to me that the scheme John mentions in #168 would have a similar
effect -- putting all the participation files someplace away from all the
home directories would make vandalism less likely to trash them, right? --
but without discriminating between long-time conferencers and everyone
else.  Then again, I may be wrong. 


#174 of 292 by jazz on Mon Nov 8 12:44:42 1999:

        Well, never attribute to malicious hackers what can easily be explained
by thousands of bored "software engineers" and other twits.


#175 of 292 by richard on Mon Nov 8 18:08:05 1999:

perhaps staff should identify these /a drive vandals and notify their
ISP's and block'em if they are all coming from the same place 


#176 of 292 by tpryan on Mon Nov 8 20:02:06 1999:

        Or put a hard limit on the amount of disk space used by a 
user?


#177 of 292 by pfv on Mon Nov 8 20:18:10 1999:

Gee.. what happened to the grex "we have no problem[issue]" speeches?


#178 of 292 by mcnally on Mon Nov 8 20:28:28 1999:

  They vanished when you woke up..  Nobody's ever claimed that we don't
  have problems on Grex.  The fact that we don't all agree with your
  approach to disk-space management doesn't mean that everyone thinks that
  the situation can't be improved..

  re #176:  The problem with enabling quotas is largely one of computational
  power -- the extra overhead for filesystem calls with quotas enabled on
  user filesystems would probably swamp Grex's fairly feeble processor(s).


#179 of 292 by pfv on Mon Nov 8 20:30:42 1999:

Aww, that was cute..

Yeah, sorry: quota is too CPU-intensive.


#180 of 292 by drew on Mon Nov 8 21:13:33 1999:

I agree with Richard on this. ISP accounts are still mostly traceable to the
perpetrators, and any ISP with half a brain is going to take a dim view of
this sort of nonsense.

As for a separate partition, this can be resolved fairly, maybe, with a 30
day? 60 day? wait before being moved to the stable partition. Certainly worth
looking at. Clean up /a, then have all new arrivals start out on /b, and later
get moved to /d (or whatever name is available) when we get to know them
better.

As for quotas: has this been experimented with to see just how bad the CPU
hit would be? Perhaps a CPU upgrade is in order.


#181 of 292 by jazz on Mon Nov 8 23:51:15 1999:

        You're assuming that ISPs are always responsible for their users. 
That's not always the case.  I've had very bad luck contacting certain
administrations (as a system admin on other systems) - occasionally they'll
simply state that they're just not responsible for what happens with a wingate
proxy server on their network, since they do not directly administrate them.
M-Net's approach was to ban wingates when they were recognized (they have a
distinctive login prompt).


#182 of 292 by scg on Tue Nov 9 00:42:17 1999:

My impression is that the disk space fill-ups are generally a lot of people
doing various stupid but probably not malicious things, that add up.  When
we get repeats of that sort of thing, from people who persist after being told
to stop, we often do follow up with the ISPs.  Sometimes it helps, sometimes
it doesn't.  However, the amount of work involved in tracking down the ISP
of every eggdropping would be pretty staggering, and generally wouldn't help
anything.


#183 of 292 by mdw on Tue Nov 9 01:20:32 1999:

Actually, the simplest approach is to add more disk space.  The fact
this is becoming more & more of a problem is almost certainly an
indication that we're way closer to capacity than we should be.


#184 of 292 by pfv on Thu Nov 11 15:04:28 1999:

        What was with the downtime & reboot this AM?


#185 of 292 by aruba on Thu Nov 11 15:41:03 1999:

I couldn't dial in this morning - had to telnet.


#186 of 292 by pfv on Thu Nov 11 15:44:31 1999:

        Ping was refused; finger was refused; telnet was refused.

        Load has been up over 20, we now have:

 10:43am  up  1:03,  65 users,  load average: 9.44, 14.74, 13.34


#187 of 292 by albaugh on Thu Nov 11 18:23:53 1999:

Don't know that this is a "problem", but why, when I was reading the agora
"man of the century" item, at the More prompt, did bbs display:

(Fixed item 119 flags 38->18 mtime 382b073a)
(Fixed item 119 rcnt 54)


#188 of 292 by omni on Fri Nov 12 08:21:41 1999:

  Nice and speedy now, even at 2400. Way to go, staff.


#189 of 292 by keesan on Mon Nov 15 15:15:10 1999:

I logged on once and got NO CARRIER after some gibberish.  Second login got
me to bbs with some gibberish along the way, but I can only read the first
one to five lines of anybody's response then it turns to gibberish.  Tried
both 761-3000 and 761-5041.  A new user also phoned (in Russian) to ask why
he got NO CAREER instead of e-mail.  (I can read what I am typing just fine,
and this is line six already).  Trying to read e-mail got me a line or two
of readable e-mail plus gibberish, for each e-mail.  We may have to go out
and fix the NO CAREER problem for the Russians if it is not a grex problem,
so would like to know before this evening if this is a general problem.


#190 of 292 by keesan on Mon Nov 15 15:16:42 1999:

Tried to read my own previous response, read 1 line, the rest gibberish.  Lynx
is 100% gibberish except for Do you really want to quit?  (I did).  I am
dialled in at 761-5041 right now.


#191 of 292 by mcnally on Mon Nov 15 18:24:32 1999:

  I keep having NO CAREER problems, too, but I've been doing on-campus
  interviews, so hopefully they'll go away..


#192 of 292 by pfv on Mon Nov 15 20:56:05 1999:

reboot NOW, avoid the rush.. 

Something is definitely WRONG when folks can leave party and we get NO
MESSAGE in the log....


#193 of 292 by krj on Mon Nov 15 21:32:04 1999:

That can be done intentionally by judicious process killing, I believe.


#194 of 292 by kaplan on Tue Nov 16 00:29:31 1999:

keesan,  the "gibberish" that you're talking about sound like a noise on 
a phone line with a non-error detecting modem.  If I'm not mistaken, any 
modem speed 9600 or above will detect such gibberish, drop it, and ask 
the sending modem to transmit again.  So noisy phone lines make things 
slow down but not fill with gibberish.  Are you using a modem speed 2400 
or below?  If so, and given the popularity of 53,000 speed modems, I 
would expect that 9600, 14,400, and even 28,800 speed modems would be 
easy to find very cheap on the used market.

If you are using an error detecting modem, then maybe the problem is the 
cable connecting the modem to the computer, the modem's power supply, 
bad modem interface, bad motherboard, or something like that.  I very 
much doubt that it's a problem with Grex.


#195 of 292 by tpryan on Tue Nov 16 03:31:10 1999:

        Kiwanis has a desperate need for external 14.4 baud modems
and internal modems of all speeds.
        So dear user, if you or your company has some surplus, or soon
to be surplus, try to get them reused by getting them to Kiwanis, 
Ann Arbor.


#196 of 292 by beckham on Tue Nov 16 13:06:03 1999:

whats it...?


#197 of 292 by mcnally on Tue Nov 16 17:18:13 1999:

  Kiwanis is a charitable organization.  Several Grexers are active
  in the Ann Arbor chapter, which (among many other things) refurbishes
  old computer systems so people who couldn't ordinarily afford a new
  computer can still get a functioning machine suitable for at least
  basic word processing and e-mail..



#198 of 292 by russ on Wed Nov 17 01:18:45 1999:

I am currently on a tty which apparently has defective flow control
to the modem.  If this is the case, it would explain Sindi's problem
too.  The defective flow control screws up sz, so I am going to have
to break my files into pieces that won't overflow the modem buffer
before they encounter a software-enforced break in transmission.

Oh, until it screws up, it runs *really* fast.


#199 of 292 by gelinas on Thu Nov 18 06:34:55 1999:

Every now and again, I get things like:

Ok: che

new resp items  conference

-->   0    0 agora  (where you currently are!)
      3    0 coop
      0    0 books
      0    0 classicalmusic
      0    0 scifi
      0    0 micros
      0    0 diy
      0    0 rialto
chemie (4)
          </UL>
               <UL>Uni courses: 3. Fachsemester - winter 1952/53
                                                                <LI>605/606
Prof
.Frau Moufang: Analytische Geometrie II (5)
                                           <LI>609 Dr.Burger: Einf&uuml;hrung
in
 die Funktionentheorie (4)
                          <LI>646 Prof.D&auml;nzer: Physikalisches Praktikum
f&u
uml;r Physiker, Teil
                    II
                      (6)
                         <LI>661 Dr.Haase: Photographisches Praktikum (3)

</UL>

<P
>August 10 - 15:
                Basel-St.Chrischona, V.Mennonite World Conference in der
                Schweiz.
 <P>I get a ride with the MCC people to Basel, Harold S.Bender rules as a
                                                                        benign
pope. Peter J. Dyck

After the name, things just froze until the window went away.


#200 of 292 by bdh3 on Thu Nov 18 08:37:11 1999:

What you are seeing is overloaded voice swiches thrashing as they deal
with an enviroment of lots of data users over voice lines. Your 'onramp'
to the 'information superhighway' is a dirt road and most 'data' users
are driving lamborghinis. Get used to it.  Or buy an ISDN or xDSL line.


#201 of 292 by mdw on Thu Nov 18 09:08:17 1999:

I don't think so -- crosstalk problems with an analog connection are
just going to hose TCP performance, but with all the error correction
and detection at the hardware level, not to mention the complexity of IP
and TCP, it's just not at all likely that a noise phone line is going to
inject data into a telnet session in quite the way Joe reports here.

It does appear that something has managed to "inject" a tcp packet into
his telnet session which is certainly rather odd, to say the least.
Since it looks extremely random, it's probably a software glitch
somewhere, and not a purposeful mistake.  It might be worth collecting
the data from multiple occurrences and seeing if there's any pattern in
them that would make it possible to determine where they come from.  It
would also be worth doing a web search to see if the text can be found;
if it can, determing the offset and size of that text within the html
might provide some interesting clues.  It might also be worth asking the
operator of that site if they have log records indicating when that page
was fetched, and if so, where it was fetched during a 5 minute window
surrounding Joe's hung session.  If you can find out where that data was
*supposed* to go, then it should be easy to figure out where it got
misdirected.

Now, the places where the data could have gone astray include:
 (1) grex -- but nobody else has reported this problem, so grex is not
        likely
 (2) the michnet dialups - a bug in the portmaster that copies other
        people's data into your session?

Joe in fact appears to use mostly michnet dialups, plus some sessions
from a desktop machine at ITD, so--
 (3) something at Joe's end -- if he's sharing his PPP connection with
        something else, like a NAT, or a web browser, or something,
        perhaps whatever manages that PPP connection has problems.
 (4) some sort of really weird bug in Joe's hardware or software --
        perhaps it's randomly inserting unused junk into buffers, or
        the tcp stack is buggy, or something.

If the problem only happens on dialup connections, and Joe can't account
for this data within his household, then I imagine the michnet folks
would be very interested in this bug.


#202 of 292 by bdh3 on Thu Nov 18 09:29:48 1999:

Perhaps so.  But I can assure you that in the heart of 'Silicon Valley'
I cannot get much more than a 28.8 connection (from everything from the
Hyatt to a 'single point' in a house while I can get at least a 44.6
connection from my house in south chicagoland with the same exact
equipment.  (From SF proper down the CALTRAIN some number of tens of
miles I have noted the same.)  Thus I conclude that the more 'data'
connections over a 'phone system' the more likelyhood of shitty
performance.  Voice telephone 'switches' rely on the concept of 'pauses'
where somebody breaths and somebody else considers a response.  The fact
is with 'modem' communication over 'voice' lines there are no such
things thus while a 'switch' may be able to 'handle' 100000 people
speaking over the same set of copper (corroded and 'dirty') wires it may
only be able to handle considerably less number of 'data' conversations.

Don't rule out a funamental inability of the 'infrastructure'.


#203 of 292 by mdw on Thu Nov 18 09:43:29 1999:

I don't think you have a large enough sample size to draw any
conclusions as to switch capacity vs. communication speed.  For all you
know, chicago may use copper from a different vendor, or may not be
faced with a phone system that is expanded as rapidly so can plan their
trunk capacity better and not be pushing their circuits to the limit.  I
believe statistical multiplexors (which *do* rely on the gaps in voice
communications) are pretty scarce now -- and a good thing too -- because
as I understand it, 28.8K modems don't work *at all* on such lines.

In any event, Joe's symptom just doesn't fit with this.  A problem with
the analog link could cause garbled data, long pauses, or actual lost
data.  Any of these will show up in ip as dropped, delayed, or slowed
packets, and Joe would be complaining a lot about "network lag".
Instead, he's looking at a logical error such as the sort PPP might make
if its header compression algorithm got confused and sent a packet off
to the wrong destination.  Since PPP strips off and regenerates the TCP
sequence number in doing header compression this is actually *very*
likely.


#204 of 292 by mdw on Thu Nov 18 09:45:41 1999:

(Which reminds me -- something else Joe might try is to turn off header
compression.)


#205 of 292 by aruba on Thu Nov 18 15:25:47 1999:

Grex just hung up on me.


#206 of 292 by jazz on Thu Nov 18 16:36:33 1999:

        I'm confused by Beady's remarks, as well.  Almost all core switch
traffic is entirely digital, and has been for some time - a DS-0 is a DS-0
whether you're screaming your lungs off at a rock concert or whether it's
completely silent, and is only re-switched to a different path when there is
a network event.

        This really is a fascinating problem.  My bet would be with whatever
PPP server Merit is using (are they still Livingston Portmasters?) becoming
confused when routing or reassembling packet fragments.  But I've never seen
anything like that from a Portmaster.


#207 of 292 by gull on Fri Nov 19 01:41:14 1999:

Re #202: I bet it's got more to do with the local copper...or even modem
incompatibilities.  Back when Michnet still had an 800 dialin, the best I
could get from Alma to the *local* Alma dialup was 21,600 baud, but I'd get
a full 28,800 to the 800 number -- which you'd expect to be the noisier
connection.  The new modems they put in here at Tech are finicky and
sometimes won't talk 28.8 to my modem at all, and insist that I manually
force it to fall back to 14.4.



#208 of 292 by gelinas on Fri Nov 19 03:03:34 1999:

Time to stop guessing, guys.  I use ISDN from my home.  I only use a
standard modem when I am out of town, which I'm not right now.

I think Marcus is on the right track: TCP injection.  The only other
machine on this side of the ISDN modem is not on the network right now.
(If I plug it into the modem, the modem never hangs up, so that <expletive
deleted> NT machine is generating *some* kind of network traffic even
when no one is logged in to it.  Blast.)

Now, this isn't the first time I've seen this behaviour on grex.  My thought
was that it was something Backtalk was doing.  I don't remember seeing
it any where else.  Note, though, that I don't use Backtalk.  (That *is*
the name of the web interface to Picospan, isn't it?)

So far as I know Merit is still using Portmasters.  However, I think I
would have seen complaints about this in ITD's Online Consulting queue,
or seen other discussion of it, if it were happening to other users.
Just as you (and i) would expect other grex users to report it here,
if they were seeing it.

Tell you what.  I'll start dumping the injections to a text file to
report here.  I'll also work through UM to see if Merit is seeing other
complaints about this kind of behaviour.


#209 of 292 by mdw on Fri Nov 19 03:34:55 1999:

It's true Jan Wolter (one of the authors of backtalk) does have a german
background, and has worked in academia, but I still doubt a bug in his
code is going to somehow go feral and start generating web pages
containing german class lists.  Backtalk is, in any event, an ordinary
cgi script - it doesn't have the ability to hack into the kernel and
munge another processes tcp connection - or even its own (apache is
between it and the outside world.)

Some ISDN connections still use header compression.  If you've got knobs
(virtual or otherwise) to twiddle on your end, it's certainly worth
experimenting with them.


#210 of 292 by gelinas on Fri Nov 19 04:28:54 1999:

No knobs.  Drat.

Well, you probably know the conferences here better than I do; I'm just a
newcomer.  I've no reason to think the quoted text wouldn't appear in a
response here.  Apparently, you do.


#211 of 292 by scg on Fri Nov 19 04:48:45 1999:

What sort of ISDN equipment are you using on your end?  Are you using NAT?


#212 of 292 by gelinas on Fri Nov 19 05:03:50 1999:

It's a 3Com OfficeConnect LAN Modem.  Yes, I'm using NAT, and allowing
"Automatic Incoming NAT".


#213 of 292 by goose on Fri Nov 19 05:44:01 1999:

I just had a really wierd thing happen, maybe related, maybe not.  After
waiting for a free port after it got down to 5 to go this appeared:
Nov 19 00:21 Nov 19 00:21 40977 -1 23733 203.246.133.92 LOST HEAD
Then when the port was free it only printed "Grex" and when I pressed
enter or any other key a few more characters would appear until the standard
grex login: prompt was in full view.  But it wouldn't take my username it only
displayed anohter 'login:" after every keystroke for dozens of prompts.
finally it took my username, but not my password.  I disconnected using my
terminal program, and reconnected only to find that it connected me with this
already 'open' port.  ad I typed my password correctly and then disconected
someone could have aped my account.  What up?


#214 of 292 by n8nxf on Fri Nov 19 12:04:58 1999:

I had that happen to me too, a month or so ago.


#215 of 292 by scott on Fri Nov 19 13:27:48 1999:

Extremely slow net connection?  


#216 of 292 by aruba on Fri Nov 19 14:50:00 1999:

Re #213: I've had that happen many times with Kermit for DOS.  (STeve calls it
"tty constipation".)  When I used to start up Kermit, something weird about my
system made it give me an error message about half the time.  I can't remember
what it said, exactly, but it was something about not being able to interface
happily with the modem.

If I then went ahead and tried to connect to Grex, the terminal would be
constipated.  I learned, after a few of those, to watch for the message when
Kermit started, and if I saw it, exit right away and start it again.  It
almost always wasn't there on the second try, and then everything worked fine.


#217 of 292 by keesan on Fri Nov 19 17:08:57 1999:

I got gibberish again today (on my 14.4 modem) and then a disconnect.  I may
have fixed the problem by unplugging the power from my fax-phone switch for
10 seconds.  Last time the problem occurred at both 14.4 and 2400 connect
speeds and usually my setup works perfectly  Despite having five phone lines
plugged into my switch (one through a 1-2 adaptor).  Jim did something fancy
with the system to make it work with extension phones, fax, phone, modem,
answering machine, and maybe it gets overloaded occasionally.


#218 of 292 by jazz on Fri Nov 19 21:47:36 1999:

        I have to wonder where the NAT is occuring that's mentioned in #212
- if I'm reading correctly, the ISDN CODEC, which has either a DHCP-assigned
or static address associated with the point-to-point bonded ISDN connection
(mlPPP, I'd wager) is performing NAT on behalf of several hosts attatched to
it, which may either have dynamically assigned or statically assigned
addresses (in the RFC 1918 range), *but* that there is only one functioning
host attatched to the CODEC.

        I thought about the possibility of a program associated with a tty
failing to die gracefully upon recieving a SIGHUP and spewing output to the
terminal after the original owner had logged off of the tty, but that'd be
more likely to be processed HTML via lynx than unprocessed HTML source.


#219 of 292 by mdw on Fri Nov 19 22:21:14 1999:

Extra output from a background process also wouldn't cause the telnet
session to "hang".  In fact, however, there are at least 2 reasons why
this shouldn't happen -- (1) telnetd can't open a master pty if the
slave side is still "in use" by the previous owner, (2) when telnetd
runs, it calls "revoke" on the line which should mark any existing
descriptors "non-read/write".


#220 of 292 by gelinas on Sat Nov 20 03:46:59 1999:

Hmmm.   I don't understand the reference to the CODEC.  The IP address
is assigned to the port on the Network Access Server, in this case a
Portmaster (I *think* Merit is still using Portmasters exclusively).
(And yes, Multi-link PPP is a possibility; I've configured my modem to
establish a second link when network usage demands it.  It's been a while
since I got the link that busy.  I usually see it when I insist on continuing
to work with telnet and the web while FTPing files.)

All packets come into the ISDN modem for the same address.  The modem then
figures out which host on my side is supposed to get them.  And then the
host has to figure out which window to display them in.

Oh, and yes, I've assigned static addresses to the machines on my side of
the network.  (For instance, this Mac is currently 192.168.1.2.)  But as
noted, only one machine is actually on the network right now.


#221 of 292 by senna on Sat Nov 20 06:45:03 1999:

How do you change the spacebar prompt settings in party?  A user was 
trying to change the menu and ended up getting stuck with having to 
space to talk, just like it used to be.  They were quite frustrated.


#222 of 292 by scg on Sat Nov 20 07:44:20 1999:

 :set space and :set nospace.


#223 of 292 by mdw on Sat Nov 20 08:02:04 1999:

I think what you are calling an "ISDN modem" is what jazz is calling an
"ISDN CODEC".  It's definitely not really a modem -- they do make ISDN
modems, but they're most commonly used for the ISP side of 56K
connections (56K requires that one side of the connection be digital).
I don't think I'd call it a "CODEC" either, but I don't have a better
name for the box you evidently have at home which is doing the IP
address translation.  In any event, that box is certainly the prime
suspect, quite possibly on conjunction with the portmaster (or whatever
it is) that's at the other end.


#224 of 292 by bmoran on Sat Nov 20 12:42:29 1999:

That very same type of screen showed up in several responses in the
rec.food.coffee newsgroup a few days ago. Might have been a bigger problem
than just a few notes on grex.


#225 of 292 by jazz on Sat Nov 20 13:54:01 1999:

        Yup, I was using "ISDN CODEC" for what most people mean when they're
talking about "ISDN modems" - the box that sits between the ISDN line jack,
acts as a terminator, and may or may not route packets or manage PPP or mlPPP.
In this case it's clearly acting as a router if it's doing NAT, so "ISDN
router" would be my term of choice, but I didn't want to rule out the
possibility it was a fairly dumb serial device like a Motorola Bitsurfr.

        I was thinking the ISDN router would be the likely culprit, but now
I'm wondering, since it doesn't seem likely that the router would recieve the
packet in question under normal circumstances - since there was only one box
sitting behind the router at the time the problem was seen - and it would be
unlikely that a third party could produce a packet that accidentally matched
a source/destination port pair and sequence number for a telnet session, and
that still wouldn't explain the connection hanging, something which I hadn't
read correctly before (and therefore didn't take into account).


#226 of 292 by gull on Sat Nov 20 18:41:32 1999:

I've always heard it referred to as an "ISDN pipeline," but that might be
some company's proprietary name for them.


#227 of 292 by scg on Sat Nov 20 20:37:07 1999:

Pipeline is Ascend's name for its routers.


#228 of 292 by flem on Sat Nov 20 20:39:24 1999:

I think the technical term is "ISDN thingy".  :)


#229 of 292 by gelinas on Sat Nov 20 22:38:51 1999:

ISDN router works.  I know it's not really a modem, since modems convert
from digital to analog and back, but that's what 3Com calls it, probably
to avoid confusing J Random Dialup User with too much accuracy. ;)


#230 of 292 by mdw on Sun Nov 21 08:13:13 1999:

Grex wouldn't know anything at all about the ISDN router.  TCP header
compression works by taking all the bits that are easily predictable
(like the IP header, tcp sequence number, etc.), & throwing them out.
They're put back into the packet by the receiving end.  As long as the
two ends don't get confused, there shouldn't be any problems doing this.
However, in Joe's case, one end is also handling ISDN connections from N
other users, and Joe's end is doing NAT translation.  So both ends have
at least the opportunity to get confused, and send packet data, with
forged "apparently real" tcp sequence numbers out.  These forged packets
will contain someone else's data, which will explain Joe's confusion,
and since the real sender doesn't know this packet was sent, the real
sender's tcp sequence number will be behind the client's sequence
number.  So, when the real sender sends data to the client, the client
will just toss it as "old" data.  If Joe can, typing blindly, convince
the other end to send enough data, he *might* be able to get it to
resync, but that would definitely be a pretty desperate measure.


#231 of 292 by gelinas on Sun Nov 21 23:37:21 1999:

It just happened again:

be played on the 3d: It would interfere with Monday Night Football.
Or be confused with it.)  So again not a good day for collegs
~
~
~
~
~
~                            than   ho    rCcg.


Note the tildes.  My cursor jumped from trying to enter the final 'e' in
college to a spot six lines down, with random text added as displayed.

I wonder what would have happened had I been typing in one of the other
telnet windows I had open at the time?

As before, the session hung until Grex closed it.


#232 of 292 by gelinas on Mon Nov 22 01:18:12 1999:

Some additional comments:

I've been using my ISDN line since last November or December.  The only 
thing that has changed is that I now use grex instead of confer.itd.umich.edu
for my conferencing.  Selective memory may be in play, but I really don't
remember seeing this kind of problem before I started using grex.


#233 of 292 by drew on Mon Nov 22 21:20:30 1999:

I'm getting GOT errors and trashed participation files again.


#234 of 292 by krj on Mon Nov 22 21:32:31 1999:

Out of disk space again:
 
/dev/sd2h             842574  748646    9671    99%    /c
/dev/sd1c            1944365 1749411     518   100%    /a


#235 of 292 by davel on Tue Nov 23 14:07:13 1999:

Re 231: Looks like part of someone's vi session, somewhere.


#236 of 292 by pfv on Tue Nov 23 18:38:01 1999:

Been an interesting morning - verio.net has twice fubared the routing into
grex.. Here's the latest "wtf?":

 traceroute grex.org
traceroute to grex.org (204.212.46.130), 30 hops max, 40 byte packets
 1  as1-nmc.gylr.mi.voyager.net (216.93.16.131)  166.663 ms  169.280 ms
169.717 ms
 2  gld-router-pool.freeway.net (208.137.20.254)  169.105 ms  169.477 ms
159.741 ms
 3  216.93.15.61 (216.93.15.61)  179.115 ms  259.464 ms  179.715 ms
 4  rtr1-ethX-X.mnpl.mi.voyager.net (209.153.156.65)  179.366 ms  249.469
ms  289.710 ms
 5  rtr1-ser3-0-9-0.lnng.mi.voyager.net (209.153.151.17)  329.174 ms
289.448 ms  269.706 ms
 6  border1-hssi16-0.arb.qual.net (198.87.206.1)  299.168 ms  289.464 ms
279.710 ms
 7  border1-hssi3-0.dtw.qual.net (131.103.4.5)  289.175 ms  189.480 ms
209.693 ms
 8  serial1-1.csr0.cv.oh.verio.net (131.103.4.18)  209.181 ms  219.471 ms
199.705 ms
 9  serial1-0.csr8.or.il.verio.net (131.103.5.1)  219.175 ms  219.448 ms
229.729 ms
10  ord0-0.qualnet-0.verio.net (129.250.16.78)  249.164 ms  239.469 ms
319.709 ms
11  ord0-0.qualnet-0.verio.net (129.250.16.78)  269.146 ms  239.462 ms
229.765 ms
12  ord0-0.qualnet-0.verio.net (129.250.16.78)  289.090 ms  289.485 ms
259.712 ms
13  ord0-0.qualnet-0.verio.net (129.250.16.78)  309.144 ms  289.478 ms
359.696 ms
14  ord0-0.qualnet-0.verio.net (129.250.16.78)  299.142 ms  339.465 ms *
15  ord0-0.qualnet-0.verio.net (129.250.16.78)  399.525 ms  399.464 ms
349.706 ms
16  * ord0-0.qualnet-0.verio.net (129.250.16.78)  369.888 ms  449.500 ms
17  * ord0-0.qualnet-0.verio.net (129.250.16.78)  509.894 ms  449.465 ms
18  * ord0-0.qualnet-0.verio.net (129.250.16.78)  489.888 ms *
19  * * grex.cyberspace.org (204.212.46.130)  409.898 ms



#237 of 292 by gelinas on Wed Nov 24 01:43:38 1999:

re: #235: Yes, #231 records _my_ vi session, up until the s in "collegs" and
following, which was popped in from . . . somewhere.  It is another example
of a recurring problem that I believe to be limited to grex.


#238 of 292 by gelinas on Wed Nov 24 03:16:32 1999:

OK, folks, here it is:

>Back to vans.  It looks like the choices are 1999 Dodge 1500 (Transedan or
>Sapphire) and 1999 Cheverolet Express 1500.  We've taken both for
>familiarisation drives (not extensiv added an HTML version of the judge's
rulin
g.  It's at

                  <a
href="http://www.cyberspace.org/lawsuit/injunction.html">ht
tp://www.cyberspace.org/lawsuit/injunction.html</a>
                                                   </PRE><br clear=all><hr>

#267
of 300 by Jan Wolter (<a href=/cgi-bin/bt/vanilla/bio?login=janc>janc</a>) on
Fr
i Jul 30 00:44:57 1999:<p>
                          <PRE>A couple notes on where things stand and where
th
ey are going, as I
                  understand it.

                                What we have now is a "preliminary
injunction".
 This prevents
              enforcement of the law until further rulings.  It is *not* a
rulin
g that
      the law is unconstitutional.  It is just a ruling that says that the

likeli
hood of the law being found unconstitutional is high enough that
                                                                people
shouldn't
 be prosecuted under it until it has been examined
                                                  further.

                                                          The ruling is quite
sa
tisfactory - Judge Tarnow appears to have pretty
                                                much accepted all the ACLU's
rea
soning, and even added some arguments
                                     against the law that they didn't make.


Thing
s do not end here.  Lots of different things could happen next, but
                                                                   the most
like
ly



#239 of 292 by gelinas on Wed Nov 24 03:22:41 1999:

And it happened again:

>Cutting and pasting from window to another caused re-wrapping, but it iss
>QUstand that part about "there are still some males who believe
                                                                that lesbians..
." - could you explain what you meant there, Sonja?
                                                   </PRE><br clear=all><hr>
                                                                           #6
                                       of
 12 by Sweet Pea (<a href=/cgi-bin/bt/vanilla/bio?login=simcha>simcha</a>) on
 Tu
e Dec 27 08:14:31 1994:<p>
                          <PRE>Wow, are there a lot of women out there who must
be going out (read:
                    sleeping with) the wrong men!  I have never had trouble
                    with
 my "civil
          rights" in bed!  Im very mainstream,a nd believe that if you make
          your

bed, you lie in it for any o

} I took time to rewrap the above to match what appeared on my screen.  In
} at least the first instance, it is very clear that it is something from
} grex.  I suspect the second is also recognisable (by some here) as being
} from grex.


#240 of 292 by scg on Wed Nov 24 05:06:51 1999:

Hmm... So it looks like it's either Grex or Grex's routers.  I just turned
off header compression on Grex's ISDN link, and also rebooted the Pipeline
on Grex's end of it, since it seemed generally confused.  Let me know if the
problem goes away.


#241 of 292 by gelinas on Wed Nov 24 05:10:43 1999:

Actually, I can only let you know when it happens.  ;)


#242 of 292 by tsty on Thu Nov 25 08:43:14 1999:

filled again ...
  
grex.cyberspace.org% df
Filesystem            kbytes    used   avail capacity  Mounted on
/dev/sd0a             109823   71181   27660    72%    /
/dev/sd0d             156783  131196    9909    93%    /usr
/dev/sd0e             706783  625481   10624    98%    /usr/local
/dev/sd0f             471183  211699  212366    50%    /x
/dev/sd2e             699223  565503   63798    90%    /var
/dev/sd4h            1952573  861289  896027    49%    /var/spool/mail
/dev/sd0h             284215  229821   25973    90%    /bbs
/dev/sd2a              31023   12073   15848    43%    /rootbak
/dev/sd2d              31023   12871   15050    46%    /suidbin
/dev/sd2f              62863     588   55989     1%    /tmp
/dev/sd2h             842574  750399    7918    99%    /c
/dev/sd1c            1944365 1720330   29599    98%    /a
g
  


#243 of 292 by otaking on Sun Nov 28 18:23:02 1999:

When I tried to access my inbox in pine, I got an error message telling me
that no such file existed.


#244 of 292 by sno on Wed Dec 1 03:36:32 1999:

I've finally figured out the problems I've been having with backspace on
grex.  Frankly, I'm surprised that the basic configuration inquiry does
not automatically create a default file that simply includes the line:

tset -Q -e<backspace>

whatever that character might be.  Since the newuser scripts *do* ask
what to use for backspace, it would seem appropriate to modify the
users .profile or .login to include this line.

The difficulty stems from the fact that a simple stty erase <backspace>
does not propagate to bbs.  It's beyond me why.  bbs fails to 
inherit TTY settings within its environment and I've been fighting with
it ever since.  So tonight I bashed and brooded over it until I came
to this solution.  Perhaps my effort will help other Linux users be
able to use the grex system with less difficulty.



#245 of 292 by gelinas on Wed Dec 1 04:12:39 1999:

Hmm... My .login says:

if ($?prompt) then
        stty intr '^C' kill '^U' erase '^H'

I don't know why it chose ^H, but that's the right choice for me.  Maybe
because of the check newuser did?


#246 of 292 by mcnally on Wed Dec 1 05:32:55 1999:

  You (gelinas) use csh and sno apparently uses bash (an sh variant..)


#247 of 292 by gelinas on Wed Dec 1 05:50:45 1999:

That's what he meant by 'bashed'. :/


#248 of 292 by cmcgee on Fri Dec 3 02:40:24 1999:

Hope this reboot solves the "few seconds to connect" problem.  It's been
taking more than 80 seconds.


#249 of 292 by orinoco on Fri Dec 3 04:36:31 1999:

it went much faster for me this time than it has been going recently.  


#250 of 292 by sno on Fri Dec 3 17:38:41 1999:

I just experienced an annoying flaw in the idle time manager.  While
composing a response offline, I saw both the 15 minute idle warning
and the 5 minute boot off message simultaneously.  The warning would
have been nice since I was still present and accountable and could
have easily reset the counter with a keypress.


#251 of 292 by orinoco on Fri Dec 3 18:14:23 1999:

....ooop, and now it's dragging again.


#252 of 292 by gelinas on Sat Dec 4 06:09:34 1999:

} Respond, pass, forget, quit, or ? for more options? !df .
} Filesystem            kbytes    used   avail capacity  Mounted on
} /dev/sd1c            1944365 1749828     101   100%    /a


#253 of 292 by katie on Sat Dec 4 09:45:05 1999:

Yeah, for a while there I couldn't send mail because it was full and I
have no children. Whatever that means. Plus I couldn't enter responses
because of all the fullness.


#254 of 292 by spooked on Sat Dec 4 11:20:40 1999:

We are getting a new disk added, but STeve is busy with other committments
at the moment.  He will add it he assures staff when he's finished with his
other work (which hopefully won't be too long down the track).  Please be
patient until then (staff are well aware of the problem).


#255 of 292 by davel on Sat Dec 4 14:11:02 1999:

Somebody explain the facts of life to Katie ...
8-{)]


#256 of 292 by tpryan on Sat Dec 4 18:16:02 1999:

        I know I can do a 'ls -als|more' to review the files I have on the
system and remove some that where only meant to say a little while.  I 
know others here can do the same. But us who use bbs are less than 10%
of the system users.  Anyway to get the word out to the other users?


#257 of 292 by keesan on Sat Dec 4 20:28:39 1999:

just got no carrier immediately when I typed pine (twice), but mail
gave no problems.  From Kiwanis, where we pay for each phone call 10 c.
Why?


#258 of 292 by spooked on Sun Dec 5 02:20:38 1999:

RE: resp: 256 Just added a message to the motd.


#259 of 292 by cmcgee on Sun Dec 5 03:48:00 1999:

Ok, what is the limit I should have in my directory?  And is there any way
to compress mail folders?


#260 of 292 by mcnally on Sun Dec 5 06:03:01 1999:

  Others may have an opinion on how much disk space you "should" use;
  my only recommendation is "as little as you can, but no less.."

  As far as compressing mail folders -- most Unix mail programs won't
  read/write compressed folders directly, so you won't be able to use
  them while they're compressed, but for things you aren't actively
  using, you can compress them using the "gzip" command from within
  a Unix shell.  For example:

     grex%  gzip -v -9 old_mail.mbox
     old_mail.mbox:              50.4% -- replaced with old_mail.mbox.gz

  Which would compress the file to slightly less than half its original
  size (typically gzip compresses much better than 50% on text files,
  actually.  70-80% compression is pretty common.)

     grex%  gzip -d old_mail.mbox.gz

     or 

     grex%  gunzip old_mail.mbox.gz

  will restore your compressed file.


#261 of 292 by flem on Sun Dec 5 06:11:09 1999:

The disk space message looks pretty funny in my Backtalk window.  :)


#262 of 292 by gelinas on Sun Dec 5 06:21:59 1999:

Well "quota -v" {"!quota -v" at a bbs prompt} suggests 1 megabyte (1024
kilobytes).


#263 of 292 by spooked on Sun Dec 5 06:40:33 1999:

Yes, 1MB, though it's not dynamically enforced.

Thanks, Mike.  For naive and UNIX frigthened users, filebrowse writeen by
Valerie allows compression of files amongst its many options.


#264 of 292 by don on Sun Dec 5 18:24:42 1999:

Isn't there something you can set to have a file system-commpressed through
chattr?

Also, if I were to compress the files that I do use often and then just have
a shell program to uncompress, run the program, and then compress, would the
tradeoff between extra disk space and used computer cycles be any good?


#265 of 292 by don on Sun Dec 5 20:40:29 1999:

Another thing that could be done as a stopgap until we get a new disk mounted
is to graft some space from /tmp and add it to /a (if I recall, they are
partitions on the same disk; otherwize there's probably some not-as-easy way
to graft).


#266 of 292 by pfv on Sun Dec 5 20:42:54 1999:

        The entire point is moot, as long as the space is filled with
        the same data, for the same reasons, repeatedly.


#267 of 292 by mcnally on Sun Dec 5 21:10:17 1999:

  re #265:  unless it's going to be quite a while before anyone can
  attend to the new disk, it'd be better not to mess with the partition
  layout of currently-active partitions..  besides, it would almost
  certainly take less time to hook up the new drive, partition it,
  newfs, and mount it, than it would to back up /a, re-do /tmp and /a,
  etc..

  re #266:  you really just can't deal with the fact that staff isn't
  interested in ruling Grex with an iron fist, can you?  it seems like
  we've rehashed this over and over again:  YES, the freedom afforded
  to users here on Grex allows some abuses, and YES, I'm sure there's
  a copy or two of eggdrop source (or whatever your latest bogeyman
  happens to be..) taking up space that we could use right now.
  but for the last time, NO, a couple of useless copies of eggdrop are
  *not* the real problem here..


#268 of 292 by scott on Sun Dec 5 21:56:53 1999:

Interesting... it looks like this is one of those indirect y2k things.  Turns
out at least one of the rare staffers that knows how to add another disk to
Sun systems is too busy at his money job certifying that all systems are y2k
compliant to have enough free time to actually install a disk.


#269 of 292 by scott on Mon Dec 6 00:16:36 1999:

I just got disconnected when one of the cats kicked over the phone.  I demand
Grex staff do something about this!   ;)


#270 of 292 by mcnally on Mon Dec 6 00:25:27 1999:

  If STeve can't get to it before then, as an early Christmas present to
  Grex I'll volunteer to install a new disk once I've gotten past my current
  state of end-of-the-academic-term madness.. 

  Until I finish my final programming projects, though, I sadly can't spare
  any volunteer time.


#271 of 292 by spooked on Mon Dec 6 01:11:28 1999:

Thanks for your offer, Mike!

Thanks to STeve's wizardry, he got the disk usage down quite a bit, but still
not in a comfort zone.  Probably also assisted a little by people deleting
some personal files per motd.


#272 of 292 by cmcgee on Mon Dec 6 16:41:10 1999:

Special thank you to Grex staff (scott particularly) and users, and an
apology.  Scott took the time to tell me how to do some unix stuff so that
I now know how to download files from grex to my computer.  Other folks told
me how to compress and uncompress files.  The outcome was that I offloaded
_ALL_ my old mail folders to my own computer for long-term storage.  

But some of those were huge files, and I probably exceeded all good-manners
limits in doing so. So an apology to everyone who was slowed by my excessive
use of bandwith in service to cleaning up my excessive use of disk space.


#273 of 292 by don on Mon Dec 6 22:31:40 1999:

What do you mean by "STeve's wizardry"? What kind of things did he do to clear
up space?


#274 of 292 by spooked on Mon Dec 6 23:37:01 1999:

That's a good question. He never tells his secrets (=


#275 of 292 by mooncat on Tue Dec 7 12:19:39 1999:

Like any good magician would tell his secrets.. .<grins>



#276 of 292 by spooked on Wed Dec 8 01:39:01 1999:

hehe (=


#277 of 292 by pfv on Sat Dec 18 22:21:48 1999:

nand     ttyt7    Wed 8pm  3days     26     23  vi RequestViewMW_ADSL.cfm


Come On, sheesh..


#278 of 292 by other on Sat Dec 18 22:55:08 1999:

for those of us less aware of the subtleties, would you explain what the
significance of user nand editing a file called RequestViewMW_ADSL.cfm is?


#279 of 292 by spooked on Sat Dec 18 23:11:42 1999:

I think he's trying to point out the 3days idle figure.


#280 of 292 by mcnally on Sun Dec 19 00:46:31 1999:

  Which more than likely indicates that the pty's "stuck", not that nand
  has actually had a session open for three days..


#281 of 292 by spooked on Sun Dec 19 02:35:29 1999:

Right. I'll just try and fix that shortly.


#282 of 292 by krj on Sun Dec 19 04:08:50 1999:

/var is full, which means party is not working.


#283 of 292 by krj on Sun Dec 19 04:30:51 1999:

Now fixed.  Thanks, steve!


#284 of 292 by pfv on Sun Dec 19 15:32:16 1999:

Thanks, folks.


#285 of 292 by sno on Tue Dec 21 16:21:00 1999:

Key response time has been crappy over the link recently.  Lags of
10-20 seconds are not uncommon.



#286 of 292 by orinoco on Tue Dec 21 23:14:40 1999:

Pine has been freezing on me recently.  Usually it happens when I use one
control-something combination that requires another control-somthing
combination -- that is, I do ^W for search or ^X for send or some such, and
then if I try to do an additional control-key command like ^C to cancel either
at the prompt that comes up, my connection just freezes entirely.  


#287 of 292 by keesan on Wed Dec 22 00:23:02 1999:

happens to me, too.  I have been learning Mail.  Jim's phone bill reflects
problems with grex.  We pay for 50 calls/month and usually stay about that.
Last month was 143 calls.  He said it took ten tries to send an e-mail.  Is
there a good chance the problems will be fixed, or should we switch service?
It is 6 cents per extra call.  A stamp would be cheaper.


#288 of 292 by flem on Wed Dec 22 05:33:30 1999:

I've had the problem with Pine freezing up, but only when using 
Microsoft telnet.  Then, it happens fairly predictably.  If I use a 
different telnet client, no problems.  This may not be what you're 
experiencing, but you might give another telnet app a try.


#289 of 292 by kaplan on Wed Dec 22 21:11:54 1999:

Ugh, the telnet that comes with Windows 95 and NT 4 is so bad!  Do Windows
98 or Windows 2000 come with the same lousy telnet client?


#290 of 292 by flem on Wed Dec 22 21:15:07 1999:

98 is the same.  I haven't tried 2000.


#291 of 292 by arabella on Wed Dec 22 23:40:36 1999:

Try CRT for telnet.  I'm using it right now.  You c an download
it from the TUCOWS web site.



#292 of 292 by keesan on Thu Dec 23 19:23:46 1999:

We direct dial and Pine still freezes up.  Procomm or Bitcomm.


There are no more items selected.

You have several choices: