Since I have not seen a discussion on LISTOWN-L about this problem, and I
suspect that many of the subscribers here are not on LISTSRV-L I am posting
the following exechange with the originator of a major problem on the SAS-L
list last night.  I received over 200 copies of this file.  Because of the
number of subscribers to SAS-L on our machine, it broke sendmail because of
the large number of short mail messages arriving as part of the same
connection with our gateway, and I understand that the spool at NIH was so
clogged that they had to shut down for a while this morning.  I understand
from the notes below that Eric is aware of the problem and is fixing the
problem, but you probably need to be careful if you have files=yes defined for
your list.
 
-phil
 
 
Forwarded message:
> From <@UVVM.UVIC.CA:[log in to unmask]> Thu Apr 15 15:57:51 1993
> Date:         Thu, 15 Apr 93 13:51:08 PDT
> From: Melvin Klassen <[log in to unmask]>
> Subject:      Re: files, files, anyone have files ?
> To: "J. Philip Miller" <phil@wubios>
> In-Reply-To:  Your message of Thu, 15 Apr 1993 14:50:30 -0500 (CDT)
>
> On Thu, 15 Apr 1993 14:50:30 -0500 (CDT) you said:
> >> I sent exactly **one** copy of this file
> >> to all the subscribers of the SAS-L mailing list.
> >>
> >Melvin, I have not yet seen a discussion of this on the LISTOWN-L list yet,
> >so am trying to understand a little more of what happened.
> >
> >1) why did you send it as a file rather than as a piece of mail?
> I sent **one** file, because the list allows files.
>
> Only those users who **want** "files" received the files.
> Send a 'SET SAS-L NOFILES' command to LISTSERV,
> if you don't want to receive files.
>
> >2) was the comment about how sending a file could lead to a bitnet worm part
> >of what you sent?
> When a list allows files, then a malicious user could send a "worm",
> like the famous Internet worm, to each subscriber of SAS-L.
> If some user "receives" the file, and runs it, then "results can be
> unpredictable". :-)
>
> >3) was the effect that you observed what you were referring to in #2.
> No!  The "multiple-copy" phenomena was **quite** unexpected.
>
> For discussion, see the 'LSTSRV-L' mailing-list, or ...
> ---
> Sender:       "Forum on LISTSERV release 1.7" <[log in to unmask]>
> From:         Valdis Kletnieks <[log in to unmask]>
> Organization: Virginia Polytechnic Institute
> Subject:      Listserv 1.7F and problems with looping?
>
> It appears that there may be a bug in the handling of peered lists that
> have "Files=Yes" specified.  We have had a locally peered list loop on
> us twice when a file was punched to the list - this was a "known working"
> application that distributed our user services newsletter every week for
> the last several months - it broke when 1.7F came out.  However, being a
> weekly thing, we didn't isolate it immediately, and figured it for a fluke.
>
> Tonight, the SAS-L list did the same thing, spawning off at least 150 copies
> of file.  I've forwarded copies of the problem file at various stages of
> the loop to Eric.
>
>                                   Valdis Kletnieks
>                                   Computer Systems Engineer
>                                   Virginia Polytechnic Institute
> =========================================================================
> Date:         Wed, 14 Apr 1993 20:54:11 MST
> Sender:       "Forum on LISTSERV release 1.7" <[log in to unmask]>
> From:         Bob Kaneshige <[log in to unmask]>
> Organization: Computing Services, Arizona State University
> Subject:      Being bombarded!!
>
>
>
> Can someone stop this bombardment that is being caused by SAS-L?
>
>   Bob
>
 
 
 
--
     J. Philip Miller, Professor, Division of Biostatistics, Box 8067
         Washington University Medical School, St. Louis MO 63110
          [log in to unmask] - (314) 362-3617 [362-2694(FAX)]