>That's the point I was referring to.  What if your header says that
>there are several peers (EB0 among them), but an intermediate LISTSERV
>doesn't not mention one as peer (e.g. EB0) ?   The DEL-REQ will not be
>propagated to that server.
>In the example brought by Richard, the subscriber is located in DEARN,
>so even if you sent the req to EB0 and our server resends the req to the
>'peers' mentioned in its header, may be the server at DEARN will reject
>the req as coming from a non-peered list -if EB0 is not mentioned there
>as a peer-  ( am I wrong ? )
>
>BUT, if the server is mentioned as a 'slave' or 'listening' one,
>the poor subscriber who wanted his name deleted will still receive
>the contributions.
>
>Best regards,  Miguel
 
Surely this is a situation which should not be allowed to occur.  If one
peer cannot or will not be updated to have symmetrical links with other
peers, it should be removed.  Ideally, a peer list group ought to have
at least one "global" owner who could correct the situation.  Sites which
feel they cannot support non-local list owners ought to be willing to
maintain peered lists in a timely fashion;  if they are not, the other
sites in the peer group ought to remove the uncooperative site despite
the possible detriment to the network load.
 
--Mark