|
Grex > Coop7 > #117: mail abuse and membership | |
|
| Author |
Message |
| 25 new of 72 responses total. |
steve
|
|
response 25 of 72:
| 
Oct 26 21:34 UTC 1995 |
Thats definable. We were thinking of 1M as a start.
|
tsty
|
|
response 26 of 72:
|
Oct 28 20:17 UTC 1995 |
and perhaps with a couple levels of "alert" when certain thresholds
are reached - suggest thatthje Subject: field include ALL CAPS ALERT
andthe body have a mildly/stronger worded content. One meg sounds
like an excellant start as an upper limit. Maybe the two threshold
triggers at 450K and 700K. discuss?
|
steve
|
|
response 27 of 72:
|
Oct 28 20:40 UTC 1995 |
It could be done, but I don't think that warnings really belong
in sendmail. We could easily write something else however, and
stick it all the default rc files for shells, so it could say
something to the user if a certain limit was reached.
|
tsty
|
|
response 28 of 72:
|
Oct 30 05:56 UTC 1995 |
hmmmmmmmm, a little note in the .login file? hmmmmmmmm, fascinating idea.
|
scott
|
|
response 29 of 72:
|
Oct 30 12:23 UTC 1995 |
Hey, here's a dumb idea... How about if newuser creates a default .forward
file for each user, one that contains only a comment talking about how not
to abuse Grex forwarding? It seems like a lot of problems comes from newbies
(Grex newbies, anyway) and forwarding abuse...
|
steve
|
|
response 30 of 72:
|
Oct 30 12:58 UTC 1995 |
It couldn't be a .forward file, or whenever the poor person got
mail, staff would get bounce messages.
Something I've been thinking of is a cron job that wakes up and
looks at all the .forward files on the system, catalogs them and
other things. It could be done.
|
scott
|
|
response 31 of 72:
|
Oct 30 17:21 UTC 1995 |
Hmm.. you can't put comments into .forwards? Oh well.
|
tsty
|
|
response 32 of 72:
|
Nov 1 06:10 UTC 1995 |
re #30 - ummmmm, would looking at all the .forward files on the system
edge, perhaps, closer to "invasive" than we would want to go? Just a
neutral question, that's all.
|
sidhe
|
|
response 33 of 72:
|
Nov 2 16:11 UTC 1995 |
It could be seen as such- has this been considered?
|
lilmo
|
|
response 34 of 72:
|
Nov 3 03:15 UTC 1995 |
Is there any way to, say, log all mail in and out for one day, and get an idea
as to how much is sent in here and automatically forwarded out, w/o getting
specific logins involved?
|
mdw
|
|
response 35 of 72:
|
Nov 4 09:05 UTC 1995 |
So far as I can tell, comments should work just fine in .forward's. In
fact, I'm quite certain it works because my .forward now has a comment.
99% of normal users don't come anywhere near abusing mail. The number
of people who cause problems with mail is in fact quite small - and 1 M
of unread mail turns out to quite a bit beyond what most people care to
find in their inbox. This mechanism isn't really meant to throttle
people's use of mail, but to provide an escape valve at a point which is
way beyond reasonable use. Someone who wanted nicer warnings could
easily write themselves something to do this. I haven't bothered
because (a) I'm already having enough trouble finding time to finish
this, and (b) the sorts of people for whom this situation arises don't
seem to be very cooperative with voluntary efforts to curb mail abuse.
One of the persons concerned has managed to be sufficiently churlish
that one grex staff person now refuses to move this persons mailbox.
Perforce, this person has managed to extract favoritism from grex staff,
at the expensive of others, through simple selfishness and rudeness.
|
davel
|
|
response 36 of 72:
|
Nov 4 11:11 UTC 1995 |
Um, Marcus, it's good to know that comments are legal in .forwards, but
what might happen if a .forward consisted of nothing but a comment - which
was what was suggested? (For all of me, that might be fine, too. I
just wonder what STeve knows that I don't - like, does sendmail choke
when I goes into forwarding mode & *then* can't come up with an address?)
On the non-techie-geek issues, Marcus's comments seem to me to be mostly
on target. We occasionally see people who can't figure out how to exit
without leaving messages in their in boxes, but I also would guess that
they're seeking help long before 1MB.
|
mdw
|
|
response 37 of 72:
|
Nov 5 10:59 UTC 1995 |
I don't think it would choke - but you're right, it probably wouldn't
end up in the user's mailbox. We'd probably need to include something
\user as well. I'm not sure anything at all is actually necessary - but
if the problem we're worried about is people creating long mailing
lists, I think we could, without much effort, build logic into sendmail
to (for instance) ignore a person's .forward if they have tried to
forward mail to more than 5 outside mailing addresses.
|
lilmo
|
|
response 38 of 72:
|
Nov 5 20:45 UTC 1995 |
Is there a legitimate neeed for .forwards containing more than one address?
|
ajax
|
|
response 39 of 72:
|
Nov 5 21:38 UTC 1995 |
It's a convenient way for users to create mailing lists - you send a message
to an account with a big .forward file, and it gets resent to all the people
listed in the file.
|
popcorn
|
|
response 40 of 72:
|
Nov 6 05:47 UTC 1995 |
(Re 35: More specifically, that's me who refuses to move TS's mail.)
|
rcurl
|
|
response 41 of 72:
|
Nov 6 06:37 UTC 1995 |
I don't agree with putting limits on any *specific* use of the system,
such as use of a .forward file for a mailing list, instead of putting
limits on *general* uses, such as filespace, or total mail, or bytes
transferred, etc. At least not until we thoroughly discuss and adopt some
rules on what is "good" use and what is "bad" use.
I maintain an account for the MNAC which is used for a board mailinglist,
some record keeping, and as a repository for reference material cited in
the environment cf. The .forward now has five names, and handles an
average of a couple of messages a week. Now, what aspect of this makes it
a "bad" use of the system?
|
mdw
|
|
response 42 of 72:
|
Nov 6 10:05 UTC 1995 |
It's bad when somebody creates a .forward that contains 50 names, all of
them off-site, sends mail *from *off-site to the mailing list, and sends
several megabytes a day this way. Apparently, this is very nearly a
routine problem these days.
|
remmers
|
|
response 43 of 72:
|
Nov 6 13:01 UTC 1995 |
Re #41, last sentence: Nothing makes it a bad use of the system.
Seems like a good use of the system.
How feasible technically would it be to put a per-user quota on
outgoing mail volume?
|
tsty
|
|
response 44 of 72:
|
Nov 8 15:11 UTC 1995 |
re #35, #40. popcorn is correct, she doesn't move my mail *I* do, expecially
when it is requested and i'm not in the middle of a 350 recipient email
list explosion with many that bounce. (Grex is *not* the source machine
for that email explosion!)
Logging in to DISCOVER - SUDDENLY that all your email has been !poofed
is not a happy login. Besides, with the suggested limits stated above, i've
only gotten as high as 20% BELOW the limit. And, fwiw, +my+ personal
limit-target is 50% below the suggested limit, sometimes rising to
20% below the limit on occassions.
If someone is not logging on regularily, and all requests are ignored
adn thje mail spool grows and grows ... staff does need to "do someting."
What the steps are/would be and what thje sequence would be is left up
to Borg and Staff negotiations. Imo, it should not be a simple "my
decision, therefore ... !poof."
Btw, !poofing thje email spool makes it disappear from the spool, but it
is available in the user's directory ... the same place where the random
bootup !poofing strikes files, directories and andything else that
is on thje users' filespace. In that case ... there is *NO* recovery,
the file that is machine-poofed is truly !poofed bye-bye for ever, no
recourse.
The safety of email from random disappearance is best demonstrated in
the spool, not in home directories. Nevertheless, there are good and
suficient reasons to move file(s) off ~/spool/mail/~. Available space
is certainly a sufficient reason.
My "fix" (done by me) is to move a bunch of email to a directory,
compress it and then duplicate the compressed version with the hope that
BOTH files won't fall victim to the bootup random-poof process.
And, two compressed files of email take LESS space than one un-compressed
email file, so there is an additional benefit with disk space.
No, popcorn does not move my mail - much mor properly (imo) she emails
me and says something nice about the condition of the ~/spool/mail/~
situation and +asks+ if I can help out the problem. Of course I do at the
next available opportunity which might be in 10 minutes or 2 days, depending.
If, however, I were to be inconsiderate (as inaccurately cast above) I
would ignore polite email, grow the ~/spool/mail/~ and +then+ be subject
to the !poof powers of root(s).
Reversing the two immediately preceding paragraphs as a matter of
system policy puts the hearse before the cart, and +that+ "system," imnsho,
is a "wrong procedure."
Also, popcorn and i have developed this "system" or ours over time with
considerations in each direction. This system did not develop in one
swell foop (intended). My take is that we were working on a beta version
of how Grex "should/could" proceed - hardly "staff favoritism" nor the
added invective.
|
scott
|
|
response 45 of 72:
|
Nov 8 16:01 UTC 1995 |
I suspect TS is a rare case of somebody that gets lots of mail, but not in
a way that needs the heavy action. I think popcorn deals with people who
either get insane amounts of mail (amounts that would put TS to shame :) ),
or that hit the limit and turn out to be absentee users who just forget they
ever had such large amounts of mail coming in and can't be found to read it.
|
popcorn
|
|
response 46 of 72:
|
Nov 8 17:00 UTC 1995 |
This response has been erased.
|
srw
|
|
response 47 of 72:
|
Nov 9 06:28 UTC 1995 |
I think TS should get treated like any other user and should not be given
any special privileges. I think that rule should apply to all.
|
popcorn
|
|
response 48 of 72:
|
Nov 9 16:42 UTC 1995 |
I agree 100%.
|
davel
|
|
response 49 of 72:
|
Nov 9 22:41 UTC 1995 |
Only 100?
|