LSTOWN-L Archives

LISTSERV List Owners' Forum

LSTOWN-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Eric Thomas <[log in to unmask]>
Mon, 5 Jun 1995 23:26:43 +0200
text/plain (54 lines)
On   Mon,    5   Jun   1995    14:27:27   EST   "A.    Ralph   Papakhian"
<[log in to unmask]> said:
 
>INFO   KEYWORDS  (Appendix   B   of  the   complete  manual)   discusses
>Auto-Delete=Full-Auto      and      Auto-Delete=Manual      but      not
>Auto-Delete=Semi-Auto,  which just  happens to  be the  default in  most
>cases. (!)
 
In Full-Auto mode, LISTSERV never forwards  delivery errors that are in a
format it understands, except (if DELAY(0) is in effect) as justification
for having  kicked off a user.  In Semi-Auto mode, it  forwards errors it
was not able to process because it could not find a matching user.
 
>The Auto-Delete=Delay(n)  option is  completely incomprehensible  to me.
>What is LISTSERV doing, exactly? If the Delay is set to 4, does deletion
>occur only  if errors are recorded  on days 1, 2,  3, and 4? Or  only if
>errors occur on days 1 and 4?
 
LISTSERV  only  receives  negative acknowledgements  (bounces)  and  they
arrive when they please.  LISTSERV has no way to know  that a problem has
been solved except  by noting that bounces no longer  *seem* to be coming
in, which could also be because they are being buffered somewhere. If the
delay is set to  4 LISTSERV will allow the problem  to remain present (ie
reported) for 4 days before it takes  action. It doesn't matter if it was
reported every  day or not and  in fact this kind  of temporary permanent
errors are often buffered at a broken machine.
 
>"For instance, if Delay(7) and  a subscriber's mail regularly bounces on
>the  weekend, LISTSERV  will wait  until  [t]he next  weekend to  decide
>whether or not to delete him,  at which point the subscriber will bounce
>mail  again and  start the  process all  over. The  bottom line  is that
>LISTSERV might  as well have  gone ahead  and deleted the  subscriber as
>soon as the first bounce occurred."
>[from Appendix B of List Owner's Manual (c)1995 by L-Soft international,
>Inc. copied here without permission]
>
>"Start the process over" and "Might  as well have gone ahead and deleted
>the subscriber" seem to me to be contradictory. Or...?
 
If you have an account that is broken every weekend and you tell LISTSERV
to wait  7 days before  kicking you out, you  will have wasted  your time
because the account will break on Sat,  LISTSERV will wait 7 days, and on
the next Sat LISTSERV will see that  the account is still broken and kick
you  out. The  account  will  have been  repaired  Mon-Fri  but as  noted
LISTSERV has no way  to know that. Increasing the delay from  4 to 7 will
not have provided any additional leniency.  This phenomenon is due to the
fact that  most false permanent  errors occur during weekends  on systems
that aren't capable of running on auto-pilot.
 
The executive summary is that the default option provides best results in
most cases and probably doesn't have to be changed.
 
  Eric

ATOM RSS1 RSS2