William Westfield (BILLW@MATHOM.CISCO.COM)
Mon 28 Nov 88 11:05:10-PST
Here is how the the SRI-NIC.ARPA mailer works, in general.
There are two queues for mail delivery. Each queue handles only one
message at a time.
The first queue handles first time delivery attempts. If a host
is down, it is marked so, and further delivery attempts to that
host (say, on additional messages) are skipped. The behavior of
this queue affects the total number of messages per day that can
The second queue handles additional attempts for hosts that were
down or other temporary failures...
tcp-ip has 415 recipients at 296 different hosts. The last time I
looked at this problem, most of the mailer's time was spent trying
to open connections to hosts that were dead (as opposed to slow
connections.) Perhaps things have changed since then.
I am not worried even if the list CAN only handle 8 messages a day.
If there are consitantly that many messages on the list, I would
expect it to become digestified, or I will unsubscribe. In addition,
because of the way the first mail queue works, a second message that
arives durring the three hours it takes to process a first message
will probably not take as long.
SRI-NIC.ARPA doe NOT include the Karels/Jacobson modifications. On
the other hand, being a tops20 system, it never behaved quite as poorly
as 4.2 unix did in the first place...
This archive was generated by hypermail 2.0b3 on Thu Mar 09 2000 - 14:44:55 GMT