|
|
| Author |
Message |
valerie
|
|
System Problems
| 
Mar 22 14:49 UTC 1998 |
This item has been erased.
|
| 306 responses total. |
dang
|
|
response 1 of 306:
|
Mar 22 21:44 UTC 1998 |
The first problem. :( I've had a report that the /b shell is broken.
|
valerie
|
|
response 2 of 306:
|
Mar 22 21:58 UTC 1998 |
This response has been erased.
|
janc
|
|
response 3 of 306:
|
Mar 22 21:58 UTC 1998 |
Yeah, first it was missing, then I fixed it, but fixed it wrong, now I've
fixed it right (I hope).
One bug down ... how many to go?
|
i
|
|
response 4 of 306:
|
Mar 22 22:20 UTC 1998 |
Two (depending on how you count) for the collection:
Bad parameters near ""
message i got when i answered 'join' to the 'you are not a member of
/bbs/agora25' prompt.
'mail' can't find my mailbox (/var/spool/mail/i/z/i). Doesn't matter
right now, though, for a couple reasons that 'll' will show you....
No hurry.
|
valerie
|
|
response 5 of 306:
|
Mar 22 22:27 UTC 1998 |
This response has been erased.
|
janc
|
|
response 6 of 306:
|
Mar 22 22:48 UTC 1998 |
Yup, all one letter logins had their mail owned by root (they had to be moved
separately). This has now been fixed.
|
senna
|
|
response 7 of 306:
|
Mar 23 00:39 UTC 1998 |
Holy cow. So far the processor response is splendid.
|
scg
|
|
response 8 of 306:
|
Mar 23 00:50 UTC 1998 |
Something must be broken. Grex isn't pausing long enough for me to take a
nap between conferences. When am I supposed to sleep? ;)
|
scott
|
|
response 9 of 306:
|
Mar 23 03:14 UTC 1998 |
The "who" command seems to be slower than the "w" command. ???
|
dang
|
|
response 10 of 306:
|
Mar 23 03:18 UTC 1998 |
We are now using gnu who, which does a lookup on all the domain names. This
has the good effect that the whole name shows up, and the bad effect that it's
slow.
|
srw
|
|
response 11 of 306:
|
Mar 23 04:04 UTC 1998 |
It's really very very slow because of that. Especially when you consider
how fast everything else is.
|
steve
|
|
response 12 of 306:
|
Mar 23 05:24 UTC 1998 |
But you can always use /bin/who, which doesn't do lookups, if
you need the extra speed.
The problems with gnu who is that if the lookup for the person
doesn't find a DNS entry (something which almost never happened,
but does today because places are too cheap to get domains), it
had to expend time doing that, which is why who listings will
seem to be going nice and fast, then pause for 5 - 15 seconds
and then go on again.
|
mcnally
|
|
response 13 of 306:
|
Mar 23 05:36 UTC 1998 |
unfortunately name resolution is a relatively complicated process
that takes a non-negligible amount of time. on the bright side,
with optimal caching, each lookup only has to go out to the internet
for a full resolution once and subsequent requests get checked against
the info that's already been fetched. on the downside, that doesn't
help all that much because each user that logs in is highly likely to
come from someplace different than the hosts that previous users have
come from.
if it really proves to be a problem it might be worth tuning the
caching behavior of named to hold more reverse name records for longer..
one would probably need to cache such records for unacceptably long,
though, for it to help a great deal.
|
mcnally
|
|
response 14 of 306:
|
Mar 23 05:38 UTC 1998 |
and, as STeve points out, the big time killer with lookups is
when the resolver waits for a failed lookup to time out..
|
tpryan
|
|
response 15 of 306:
|
Mar 23 05:44 UTC 1998 |
I guess you can't do !w -n|more like on M-net where the -n
does no new lookups. just returns numbers.
tried, w: bad flag -n
Maybe a local fix?
|
senna
|
|
response 16 of 306:
|
Mar 23 06:05 UTC 1998 |
I'm still dropping some text, though
|
aruba
|
|
response 17 of 306:
|
Mar 23 06:20 UTC 1998 |
Pine's "-i" flag doesn't seem to be working; it used to take you directly to
the screen which lists the contents of your mailbox, but now it doesn't seem
to do anything.
|
mcnally
|
|
response 18 of 306:
|
Mar 23 06:25 UTC 1998 |
hmmm.. perhaps it's being overridden by a default value for the
'initial-keystroke' setting? in any case you can miming the same
behavior by going to the <M>ain menu, then <S>etup, <C>onfig, and
changing the value for 'initial-keystroke' to "i" (for <I>ndex)
|
aruba
|
|
response 19 of 306:
|
Mar 23 06:42 UTC 1998 |
Interesting - didn't know about that. THanks, Mike.
|
valerie
|
|
response 20 of 306:
|
Mar 23 15:09 UTC 1998 |
This response has been erased.
|
birdlady
|
|
response 21 of 306:
|
Mar 23 19:44 UTC 1998 |
If I don't want a list of users *and* their addys, I just type !users. It
also keeps me from having to scroll back three screens.
|
gibson
|
|
response 22 of 306:
|
Mar 23 19:54 UTC 1998 |
I didn't know about the -i flag. I tried it and it works fine but when
I used config to set it I got "bad initial keystroke (missing comma?)". What
am I doing wrong?
|
aruba
|
|
response 23 of 306:
|
Mar 24 06:25 UTC 1998 |
Thanks for the fix, Valerie.
|
keesan
|
|
response 24 of 306:
|
Mar 25 00:13 UTC 1998 |
I tried to abort (cancel) a message that I was writing using mail,
by hitting control-C twice, and what happened was that ther appeared
on
Well, the backspace is not working for me, but that may be something
else related to Jim's putting a mini version of z
zmodem on instead of procomm
The problem in mail, even before he changed to zmodem was:
^C@
^C@
appeared and there was no way to abort the message.
This happened twice already with the Ctl Ctl-C.
Also, it took 25 sec to ket
get a response to kermit -r, and a minute and a half for zmodem sz, which I
think used to come up quite a bit faster.
Please excuse any other odd things that Jim's zmodem did here, I cannot get
pico to work eiether and hope you don't see what Isee. Let's see if I can
abort this response now.
Well, I could have aborted the responses here, but not in mail.
|
