response 1 of 77:

Oct 26 21:10 UTC 1997

Regardless of the effects, we're stuck with it for the moment.  I'd say the
easiest thing to do to fix it, since we're going to do it anyway, is get the
670 up.  However, it's also likely to be the second longest of the three
options.  These are, as far as I know:
        1.  Get up on the 670.  It will work with the ALM (serial) card fine,
            and the modems can go back to what they were.  This will happen
            eventually, regardless of this problem.
        2.  Fix the telnetd, rlogind, and sshd programs to route traffic from
            the terminal server to a seperate set of ptys, thus bypassing the
            telnet queue for dialins.  This will happen regardless of the
            problem.
        3.  Fix the powersupply on Grex, so that we can go back to using the
            ALM card.  This won't happen unless this problem spawns it.

Of the three, 3 will likely be the quickes, then 1, then 2.  However, 3 is 
also expense that need not be spent, as both of the other two will happen
eventually.  So, can we stand it for however long it takes to get up on
the 670?  Or do we need it fixed now?  One other think to keep in mind:
Many staff members dial in.  This problem will likely spur them on to more
work on the 670, but working on the power supply may possibly detract from
work on the 670.  I don't know.  Keep in mind too, that no fix is likely to
be in place in less than a week, probably more.

All this said, I don't really have an opinion, because I telnet in since the
advent of teh ISDN link.  (Mostly, my telnet software is much much better than
my dialin software) I'm leaning slightly toward waiting for teh 670, but I
do know how annoying it is to dial into the queue, so I'd also like to see
this fixed asap.  Option 2 waits on Marcus, and I have no idea where he stands
on it.  I know it's on his todo list, which is a mile long, but I don't know
where.

response 3 of 77:

Oct 26 22:26 UTC 1997

I'm with dpc on this one - I'd like to see it resolved as quickly as possible.
I suggest putting a time limit on solutions other than fixing the power 
supply.  In other words, we could say that if the situation isn't take care of
in, say, three weeks, we spend whatever it takes to fix the power supply or
get a new one.  (I heard at the board meeting that it would cost $150-$200 for
a new one, which is money we have in our bank account.)

response 5 of 77:

Oct 27 04:55 UTC 1997

The thing is, we're not sure fixing the power supply will do it.  And even
when that's done, it would mean going back to the other problems associated
with running the modems through the serial ports.  Getting on the 670 isn't
a good solution either, because really we should still be using the terminal
server there.  It would be much better to get the telnetd to the point where
it can tell the difference between local and terminal server users.  I'm not
sure what Marcus's current timetable is on that.

response 7 of 77:

Oct 27 21:56 UTC 1997

I'm with steve #5

response 11 of 77:

Oct 29 23:20 UTC 1997

Ok, is there some way you techies can fix things so that there is less than
90second wait between "wait for a conection" and the start of the telnet que?
My problem seems to be in my software, that quits after 90 seconds of waiting
for the queue to start counting down.  It thinks that there is "No Connection"
, shrugs its shoulders, and goes off to do other things (pout, I do believe).

response 13 of 77:

Oct 30 00:25 UTC 1997

Same problem as Catriona - every time the numbers start, I stay connected.

response 15 of 77:

Oct 30 06:06 UTC 1997

Re: scg and jared:  In this particular case, I am very reluctant to accept the
normal "we'll do the best thing no matter how long it takes us to get around
to it."  I'm looking for something to put out the fire here, and I think we
should be willing to spend a little money to do it.

I think it's a very serious problem, making dialin users go through the
telnet queue.  Argue about the seriousness if you like, but if you agree
that it's serious then we ought to do something about it pronto.  Buying a
power supply is something we can do.  If it doesn't work, then at least we
tried, and we can sell the thing to someone else. 

response 17 of 77:

Oct 30 07:04 UTC 1997

Steve, having recently visited your place and worked with your 
computer setup, I think that you need to put yourself in the place of a 
Grex user who is somewhat less fortunate than you.  
 
Grex users who have PPP-capable setups, multiple windows, and a 
commerical ISP -- for them, the telnet queue isn't a big problem.  Stick 
the telnet queue in a window and go do something else.
 
But grex has a lot of users on heritage equipment, and we have users who 
don't have a commercial or academic ISP.  For those users -- putting 
the local dial-ins into the telnet queue is pretty much the same as 
crippling local dial-in access beginning at 11 pm - midnight every night.
 
I'm not sure what the right approach is here -- I'm somewhat partial 
towards trying to avoid work on the current system to point resources
at the Sun 670.  But I don't think we can declare this a non-problem.

response 21 of 77:

Oct 30 22:34 UTC 1997

What _is_ groupie?

response 23 of 77:

Oct 31 03:24 UTC 1997

um...