174 new of 234 responses total.
Grex is up to its old tricks of letter displacement in cooperation with lag (I'll type something out in party, the line will be lagged, and a random letter from the middle of the statement will be transposed to the beginning). Also, I'll occasionally type something fairly quickly, which will refuse to echo until I type something else later. Thus, I could type "!tel valerie," type "Hi! How are you doing?" and have it displayed nicely in party when I type the letter i or something five minutes later. These problems occur in telnet only. My only problem when I dial in is severe random text echo loss. Still :)
Interesting. I've experienced what seem to be occasional noisy lines at grex's end. (I dial, then hang up fast when i hear a really staticy ring sound, then redial and connect quickly after hearing a good, clear ring. Things occasionally go downhill fast later, though - usually the keyboard echo stops, and my modem starts flashing "retraining" - often the connection come back a few times before finally dying.)
The ring sound, on standard Ameritech POTS lines, is generated by something in the phone switch, before it actually gets onto Grex's lines. If you're hearing a really scratchy ring sound, I don't think that's a line noise problem at Grex's end.
This response has been erased.
i went through the telnet queue several times today, and each time, the session would hang and die either at 1, or 0 or right after asking for a password. it was really bizarre. i disconnected and dialed into grex directly and did not notice any severe or unusual load conditions. any ideas what happened?
That didn't happen to me, but I did log in, saw the motd etc, and then it gave me another login prompt.
Re #66: I think that happens occasionally to people whose login shell is tcsh, as yours is. My shell is tcsh also, and it happens to me from time to time. Some bug in tcsh causes it to die under certain circumstances just after logging in. I'm not sure anyone knows why. When this happens, a big-looking core file gets dumped into your home directory. Then next time you log in successfully, you should delete it.
Whowzer! 8.6 MB! Is there a sweep for core dumps? Login failures have happened now and then, but I've never seen that core before (well, not recently) - so does someone delete them routinely and I just haven't looked at my directory at the right moment?
Yup, core dumps are routinely deleted by a system process.
The ls listing made it look like an 8.6mb file, but it was actually significantly smaller -- a so-called "sparse file".
for the last couple of days ...day and a half ... from two sites i have been geting 'no route to host' for both cyberspace.org as well as grex.org. finger fails, telnet fails .... waz de prob, oh mighty ones?
What are the IP addresses of those sites? When you do a traceroute, how far does it get? If this was last night, while Grex was down, that was what I was seeing too, but you shouldn't have been seeing that while Grex was up, and Grex wasn't down for a day and a half.
I tried dialing in last night, got connected, and no login prompt.
TS-- you looking to kick a person, not a machine?
Oops, nobody announced it: Grex was down from maybe midnight[?] to 9:10am this morning.
uptime says: 11:27pm up 8 days, 13:18, 39 users, load average: 1.28, 1.43, 2.38 Was Grex up, but not accessible?
The console terminal crashed, which dropped Grex into ROM monitor. When that happens, it's possible to tell Grex to resume where it left off, rather than rebooting it.
So uptime is not a reliable source of information on how long Grex has been up?
It generally is. We should look into using a more reliable console PC.
eieio ...no, what promted that question?
oh, scg ... it was before i wnet to sleep ... and after i got up...which may not be 36 hours, by count.
Reference to a sign you posted on a particularly ornery machine once...
expound please?
The machine now lives somewhere along Platt Road...
'..a more reliable console PC': What are the requirements? is it just emulating a dumb terminal? If so, I have several ADM-3a's and an IBM 3101 that I would be most willing to donate.
Well, it's an original IBM PC, and it seems to lock up about once a month. We used to use big 60-line Ann arbor Ambassador terminals, but I think those got too flaky to use.
Some phone line may be ringing open. The last few times I've dialed in, it
rings several times, then rings differently (different rhythm, I think -
probably switching to another line), then connects.
This morning's crash garbaged my participation file. Bleah. But that was
quite a storm, and Grex was right to run & hide. 8-{)]
why was grex not up yesterday? I really want to know.
RE #88 Three words, tendo -- thunder and lightning.
so grex hid from the storm?
I added a bunch of lines to my .mailrc in an attempt to cut down on those ridiculous large headers. Typical lines I added were discard Received: discard Message-ID: Unfortunately, these didn't work. Is this a system problem or my stupidity problem?
How do you get the pine program to not list al the other recipients of e-mail? Someone just sent me a change of address, but he also sent it to 100 or so other people.
Re resp:91 - I assume you're using the standard 'mail' program to
read mail. I can't find any mention of a 'discard' command in the
man page, so it appears not to be a system problem that it doesn't
work. This leaves the alternative... :)
Something that does work here is the 'retain' command. The arguments
should be the list of headers that you *do* want to see. For example,
you could put this in your .mailrc:
retain date to from subject
and 'mail' would show you just those four headers when it displays a
message.
John, I am using the "discard" command at valerie's suggestion.
I tried it on M-Net, and it worked perfectly. However, on M-Net
I did *not* have colons; on Grex I did. Ergo, I'm going to delete
the colons.
Thanx for letting me know about the retain command!
Yep, deleting those colons did the trick! Sometimes I'm so clever I scare myself. 8-)
Oh- I did encounter a *true* system problem late yesterday afternoon. I dialed in on 761-3000, got a connection, entered my loginid and password, and was promptly disconnected. I got the same thing on 761-4931.
There was a vandal bothering us for a while then, then until a little after 6pm we had it closed to get some tracking done.
Re resp:94 - Hm... My problem was that I was looking at the 'mailx' man page, not 'mail'. The mail page documents 'discard', but the mailx page does not. Odd - I thought they were the same program.
The motd states, at present, in part: > Are you having a problem with some feature of Grex? Have you read the FAQ? > It contains answers to most common questions, and can easily be seen on the > web at http://www.cyberspace.org/faq.html or by typing the "faq" command. > (Type !faq if you are at a prompt other than a shell prompt.) -srw However, running faq merely does nothing at all. It's a link to /usr/local/grex-scripts/faq, whose contents are the following: > #!/bin/sh > # > # Display Grex FAQ sheet > # > ###menumore /usr/local/grexdoc/grex-faq > Since the menumore line is commented out, it's not surprising that it doesn't work, is it?
I have another dumb problem with Unix mail. The escape character
for mail is ~. The escape character for the editor is :. I want
to set the escape character for mail to be : also.
I inserted "set ESCAPE=: in my .mailrc, but it didn't work.
What's the right command for this?
set escape=:
Thanx, mdw!
I see someone fixed the faq script, now uses lynx. Nice!
I don't know who commented out the menumore line. It was working for me when I converted the Grex faq to text form and put it in /usr/local/grexdoc/grex-faq (that was on July 12) The lynx version of the script is better if it works for you. Users with dumb terminals will not be able to run that script, though. I guess they can always do a finger faq, if they know to do so.
Nethack gave me a "Waiting for access to record." error 60 times before failing. I assume 60 retries is the default, so that's not a problem, but is there a problem with a nethack record?
.... on a buncha items today, i have seen the item header, then a second line of <item is forgotten> .... and then a screen full of responses headed by 98 new of 102 resp. this seems to be a problem. they were not <linked item> items either. well, two were, but he other 5/6 were not. i recognized the linked items items.
Try "set forget".
A couple of bugs that should be addressed: * The mail length limiter, while it displays the message when a mail file being sent exceeds the 100K limit, sends the mail anyway. * When an item in BBS is responded to for the first time in several months or years (I believe the limit would be a year, based on how dates show up in an 'ls'), the entire item is displayed on a 'read new'. (Partially responsible for some of the huge newresponse files.)
Additional: The mail software, upon getting a large file, *both* sends the mail *and* bounces a copy - in its entirety. (I would that, if it coundn't send, that it does neither - or else truncate most of the bounce.)
....set forget has been my default for eons ... and i didn't do any of the setting of forget/noforget ....however, issuing the commands seems to have cleared the problem. oh,well.
Re 108 second complaint: Drew, that's what happens if you've used "fixseen" to make items look as though you've read them. That's not the way I would have designed it myself, but knowing Marcus I'd guess it's not a bug but a feature I don't happen to like. (Instead of using fixseen I always do "read all pass >/dev/null".)
(Er, it also results when you join a new conference, which appears to functionally set things the way fixseen does, I think. The items get marked as having nothing new, but no responses are marked as read, however that's stored internally.)
This response has been erased.
"mailx" is also the name of the ucb mail program, in older system V releases.
Why can't cron reap the old lock files?
This response has been erased.
This response has been erased.
This response has been erased.
Thank you!
Jim is trying to figure out why the neighbor, who we are trying to get to join grex, does not get anything but a 'screen that freezes up'. He programmed it to wait only sixty seconds. Does the internet connection not working have something to do with it taking a few minutes to connect to grex?
There are a total of three of us on grex at the moment, the other two are not accepting messages. I thought they might like to know the number. Do we really need 14 phone lines? My neighbor may soon be the fourth. Best of luck to staff in figuring out the problem, and thanks.
Grex is, indeed, off the net.
Hey, I'm one of those dial-in only users. Only one on now-- see last item.
When I tried to send e-mail, it refused (to some place out of state?). When my neighbor finally got on grex and sent mail to EMU, it said it was sent. Where did it go? Why did mine not go?
I am the only one logged on! Tim and Arthurp, where are you?
Again, only one logged in on dial-ins. It's taking a few minutes toget the login prompt instead of a few seconds. I can see many giving up, thinking whole system is down. I did give up earlier tonight and went and caught up on M-net.
Ditto on the several minutes wait. Something must have to time out before Grex agrees to open a connection.
Grex is off the Net. The router at the other end of our Internet connection isn't seeing its ISDN line. I've been waiting to hear from the person whose apartment it is in about whether it is unplugged before I call Ameritech, but I should porbably go ahead and call Ameritech at this point. I'm also working on the several minutes to log in problem. I expect to have it fixed shortly.
Ok, it's an ugly fix, but the delay on dialing up is gone.
I've called Ameritech to report the problem with the line. They're supposed to test it and call me back sometime soon. Given the weather, I'm guessing this is a wet cable problem, but there are any number of things it could be.
4 hours later, there are still only 4 users on - all local, of course. (jdeigert, remmers, other, & me - hey, Sindi, don't you folks ever sleep?) Steve, we really appreciate your working on this stuff all night. Pity the net link is still down ...
This feels like the Grex of six years ago, before we even *had* an internet connection. What a nostalgia rush!
Use Item 113 to tell us what Grex was like before the internet connection.
is the name thingy down?
The name server should be working for stuff that it is authoritative for, but not for anything that it needs to talk to the outside world about. I just called Ameritech back since I hadn't heard back from them. They are promising to have it fixed by 6 pm Saturday, but they say it may be sooner. I'm in the middle of some other stuff, and didn't have time to try to track things down sooner.
Thank you. That answers my question. Also why you and I are the only ones on Grex at this moment, steve.
(But Grex was never this fast before the internet connection.)
It's now past the promised 6:00 p.m. Internet line restoration time. Any idea when the darlings at Ameritech are going to get around to us?
I'm starting to need e-mail detox.
They're still telling me it's in the central office being tested. I can't get ahold of a supervisor there this time of night, and the person I talked to on the phone said she couldn't get ahold of the people at the central office this time of night. I need to do some serious screaming at them in the morning.
iwbg, iwbg .......
*What's* in the central office being tested?
our tolerance
The M-Net Board voted to create a Grex Conference on M-Net, which should be up in a day or less. FWs are: John Remmers, Steve Andre, Valerie Mates. Hopefully this conference will be used from time to time as needed, just as the M-Net Conference is here on Grex.
Figures. I picked the one week to be out of dialin range that I really need to be able to use the dialins.
Ameritech is sending somebody out to look at the line tomorrow morning.
tomorrow Monday, or tomorrow Tuesday?
Is there anything I can do to expedite the progress or lack of it? Grex is my primary, actually only e-mail aonduit and I am getting very frustrated. Let me know what I can do to get things rolling again.
We are back on the Net, as of around 6 pm today. Ameritech lost track of our trouble ticket twice, apparrently, such that it spent most of today waiting around well past the most recent due date for a technician to pick it up. I finally got ahold of somebody there who was willing to spend a lot of time on tracking this down, and had a technician paged and told that our ticket was his highest priority, and to get out there and fix it *now*. We should be opening the system back up to users as soon as the mail finishes coming in. Right now the system load has come back down to a reasonable level, but incoming mail still has the Net connection running at 100% of capacity.
Yay for all the staff work that got Ameritech to FINALLY pay attention, and to troubleshoot everything so we were sure it was them, not us!
Many thanks for the work you did to get Grex back and operating. I think I have to get a backup isp. I keep thinking about it, but this latest downtime showed me I have to do more than think about it.
Thanks, Steve G., for all your work on this.
Not only should you get a refund/reduction in charge for the days of lost service, particulary due to lost trouble ticket, you all should send them a bill for consulting on the solution.
Great work!!
What was the "root cause" of the ISDN problem? I.e. was this a "once in a blue moon" thing, totally unanticipated? Or could this situation arise again, in much the same (unexpected) way?
Could the comment flashed on the screen when a connection attempt is made and terminated be made to show longer? I have to turn on capture and then reconnect to read what it says.
re 155:
When Ameritech called me back to say it was fixed, it was a couple of
hours after the line had come back up, and I was in the middle of about five
other things that absolutely needed to be done at work. Having already spent
seeral hours on the phone with them, my goal at that point was to get them
off the phone as quickly as possible, and I didn't think to ask what they had
done to fix it. I probably should have.
You could probably call Ameritech back with the trouble ticket number and ask for the gory details.
I've already spent more time on this than I want to. If you want to call and ask, I can send you the ticket number.
I keep getting that "mkids too small" error from Picospan again. Could someone tell me what that means, again?
It means you should quit picospan and run it again. There's a bug and it runs out of space in an internal array after being used for a (longish) while... The "root" cause of the ISDN problem was probably water and bad cable. Since something very like this problem happened once before, yes, it probably will happen again.
Pardon my ignorance: Which stretch of grex's ISDN cabling is "bad" and/or subject to "water problems" ?
With Ameritech handling it? Probably most of it. :P
The part of grex's ISDN cabling that appears to be unreliable is the part from ameritech's central offices to the blackrose society.
"The Blackrose Society?"
You know, our ISP.
Oh.
So we can take some "misery loves company" comfort from knowing that all of grex's ISP's customers were affected? ;-)
Yeah, that was my question. Did *everyone* who uses our ISP lose net service for 5 days?
And what, pray tell, is "The Blackrose Society"?
Grex doesn't use a "real ISP." We have a very generous local person with very good net connectivity to his apartment, who is giving us a connection. I don't know if he has any actual customers served out of his apartment, or if anybody else besides him and us connects through there. Both of his T1 lines stayed up, so the problem was with our specific ISDN line.
(Side note, strangely relevant: The "Black Rose Society" was in an old Avengers episode...)
Side note: There was a candy that was originally supposed to have been called Black Rose. The printer misheard the candy manufacturer, and to this day you can go by a box of Black Crows.
(Note: the fault was in our telephone line going to blackrose, not *at* blackrose.)
Yesterday and today I got a bizarre result when I dialed in. Instead of getting "grex.cyberspace.org login:" I got "grex.cyberspace.org ". There was a blank space where "login:" should have been. The cursor was in the correct place, and when I typed "dpc" it took the login and all the "login:" reappeared when I entered my password.
That sounds like a problem with your computer or terminal software, not with grex.
This response has been erased.
That happens to me on occasion, I believe it's just an echo error. I've decided that my text droppage problem is local to my system, since it happens on mnet just as much.
It's a non-repeatable peculiarity. Happens to me to.
I'm running ClarisWorks for the Mac. The reason I mentioned this oddity is that I have *never* seen it before, on Grex or M-Net.
If the "login:" message came back when you entered your password, that's not grex - the login program on grex simply doesn't have the smarts to move the cursor up & rewrite the login: prompt on a previous line. It's far more likely that your terminal software remembered the "login:" message, but somehow forgot to actually write it out to your screen memory. When you entered your password, it probably had to scroll the display or do something more major, thus causing it to refresh the screen, and incidently to display the "login:" message it had remembered. Chances are that if you had done something else that would cause the local display window to be refreshed, such as temporarily placing another window in front of the terminal window, or using any local scrolling/windowing screen controls, that this would have caused the "login:" message to be displayed as well. Some terminal programs are more brain damaged than others. The stock windows terminal and telnet clients seem to be particularly brain dead. Probably MicroSoft has a corporate theory that you should be using local gui driven software, rather than remote TTY interface software. From a business perspective, it makes perfect sense.
I think the problem is that the connecting-to-terminal-server process occasionally barfs, causing modem junk to happen. Sometimes the terminal-server-to-Grex connecting process generates some junk. In either case just dialing again gives a good connection, and I've seen no data indicating any particular dialin line has a problem.
One guy, on an old obsolete cumputer with something that he doesn't understand in the way of software to connect to grex and he has problems with m-net and he says its grex's problem.... Yeah, I buy it. OJ was framed. And Clinton didn't get blowjobs from Monica. Yeah.
I'm not SunOS literate enough to answer, but that kind of delay sounds
like a wrappers problem associated with hosts that do not have working reverse
DNS (which requires DNS to time out, usually around ten seconds, but not
necessarily so) and telnet.
Another possibility is that it's got something to do with the modem
handshaking at the beginning of a session. Non-error-correcting modems often
encounter delays after beginning to train with error-correcting modems (which
GREX does have).
I dialed in this morning and read a few items in Agora. Then just now I logged in again. My problem: The responses that I read this morning were *still* new. This is the first time I've seen this happen. Howcum?
I've had that happen when I drop carrier....usually, to avoid it, I quit out of pico span <to the login prompt> and quit from there to disconnect
What toking said. You have to quit Picospan for your participation file to be saved.
In my .cfonce file, I have added the command set autosave it forces picospan to save your participation after each item. If you disconnect or get disconnected, only the last items shows up as new again. Wish I could remember what 'set edalways' is in their for. 'set stay' returns you to a respond or pass prompt after you respond (or forget).
"set edalways" causes text entry - your entry of an item or response - to bring up your designated editor instead of using Picospan's builtin text collector. For most people, these days, the editor in question is gate (thanks again to Jan!!!).
Hmm. Grex seemed pretty slow. /a/d/a/davel$ uptime 6:02am up 16 days, 16:33, 72 users, load average: 59.73, 37.22, 22.41 It's a little better now.
I dialed in, and the modem connected but after that nothing responds. Still sitting there connected to the modem after 8 minutes. (I'm telnetted in to enter this. I have problems if I telnet in, so I avoid it.)
I've had that problem too, twice, but the 761-3000 number works fine, and both times I've just tried the 761-5041 number again a few minutes later and have connected fine.
The outgoing mail limiter, upon getting a message exceeding the intended limit, *both* sends the mail anyways *and* bounces a copy back to me. I do not think that this is what's intended.
Re #186 et seq--I *did* properly log off the first time. If I'd been disconnected, I would have expected the problem.
This response has been erased.
I'll send them a note.
Grex appears to be suffering from massive processory slowdowns at the moment. Or it did, when I was getting on.
This response has been erased.
You mean, staff doesn't automatically know or figure it out as soon as they get on? :) I had assumed that was what was wrong, but it didn't occur to me to mail anyone about it.
There aren't always staff members logged on.
And to finish John's statement "but there are several staffers who are rarely, if ever, not in close contact with their e-mail."
Hmm. I'm dialed in, and it seems as though Grex is running reasonably fast, but the *display* is slow. As though at 1200 or maybe even 300 bps - displaying distinctly one character at a time, as though a fairly fast typist were typing very steadily. (I'm connected at 2400 bps, my modem max, & this is **much** slower.) When a line just wordwrapped (in gate), & when I backspace over anything, I see the very distinct backspace-space-backspace sequence. Are network packets going out to the term server with a single-character maximum size or something?
This response has been erased.
This response has been erased.
I just got in okay (14.4K) on -3000.
re 205:
I would guess that that's a modem thing, and not a terminal server
thing. The thing to do is probably to go to the Pumpkin and power cycle the
modems.
note: I don't use the dial-in modems, so I'm not very motivated to go do
that. Somebody who cares about the problem should go fix it.
This response has been erased.
I'm glad I didn't go tickle the modems. If a staffer is going to the pumkin this weekend to reboot or backup, I would like to know so I can try to be there to see it done. I don't want to have to try it going on just reading about it. My number is on here, and in the phone book.
Often when reading mail in Pine I get this banner at the top: This message contains non-ASCII text, but the iso-8859-1 font has apparently not yet been installed on this machine. (There is no directory named /usr/local/src/metamail-2.7/fonts.) What follows may be partially unreadable, but the English (ASCII) parts should still be readable. The message always displays correctly, presumably because the sender doesn't use any wacky characters. Is there any chance we could get that font installed?
Are we running X to be able to display fonts even if the font were there?
Well, no. I was just hoping that by putting the right file in the right directory we could make that message go away.
It might ... to be replaced by a "you are not running X" message. metamail can only do so much.
Could you explain what metamail is, Dave? And why Pine is running it?
Actually, I think pine may have some degree of mime support built in; I was thinking of elm, which is what I use. In the case of elm, they decided to build in (if you compile it in) calls to an external, but freely available, program to handle mime formats which elm doesn't know are basically just text. I suspect man metamail will tell you all you might wish to know, and more. Again in the case of elm, I think there's a place in the config file (.elm/elmrc) to specify fonts (or whatever they are - character sets) which elm will treat as equivalent to US-ASCII & display without calling metamail. I keep meaning to add a couple to it, but I never remember when the time comes.
Could also get your freinds to set their fancy-stanzy mailer to send you only text, instead of the rich-text-format wanting to be sent from MS mailers.
Where I see this, or 99% of it, is on a mailing list. The listadm keeps posting notices about this. I don't think anything I say will have too much effect, in my case. <sigh> Yep, everyone in the whole world uses Microsoft products to read their mail. If they don't they should. So Microsoft feels it's fine to just make all kinds of non-text formats the default.
Was just prompted with a login prompt while reading Agora. What gives?
Re #210: Either that font's not installed on the machine you're
reading it on (text or no) or metamail doesn't have the proper symlinks.
Metamail will try to read the font in, even if it won't display it in a
text-based reader.
I'm glad this situation with MIME formats came up. I use Unix mail
("!mail"). Every once in a while somebody sends me mail with a MIME
attachment. The attachment shows up as junk on my screen. Of course,
I don't know that the mail will have a mime attachment until I've
already read it! I assume the message, including the attachment,
ends up in my mbox.
Is there any way to re-read this mail with a mailer
on Grex that *does* support attachments? I lost an *important* piece
of mail once and don't want it to happen again.
pine -f ../mbox
or elm -f mbox (I'd guess from Steve's response that pine assumes your "folder" is in some subdirectory, but elm doesn't).
I thought that elm didn't handle MIME, or something like that. Yes, pine by default assumes that folders are in the mail subdirectory.
elm calls metamail when it encounters MIME-encoded messages. For that matter, you could probably call metamail directly, if you wanted. I've never done it & don't know the calling sequences off hand.
Thanx, people!!
Is agora27 ready yet?
Pretty much, I think. Katie needs to enter a first item, then some cfadm person needs to make it public. Valerie's handling that part, I think.
This response has been erased.
How come when I send an email to my grex account, and when I go to my inbox to view it, it shows an error that I cannot go to my inbox? -Saket Please reply to saket13@cyberspace.org or saket13@hotmail.com
How are you trying to "go to" your inbox? Exactly what error? (Just in case: if you're trying to access your mail via pop3, Grex doesn't support it. You need to log in & use mail or elm or <shudder> pine.)
[Dave is a hardwood kind of guy.]
mail is a hardwood?
No, but it is "hard". ;)
elm is hardwood (sigh...have to explain these things to some people.....)
You have several choices: