Date   : Sat, 30 Oct 1982 15:15:20-EDT (Sat)
From   : Keith Petersen <w8sdz@BRL>
Subject: [obrien: Re: Failed mail from Rand-Relay]

In case any of you had rejection messages from Rand-Relay...
the problem has been fixed.  Here's what happened:

----- Forwarded message # 1:

Date: Saturday, 30 Oct 1982 12:09-PDT
To: Doug Kingston <dpk at BRL>
Cc: obrien at RAND-RELAY, dcrocker at UDEL-RELAY, mike at BRL,
    w8sdz at BRL, wancho at BRL
Subject: Re:  Failed mail from Rand-Relay
In-reply-to: Your message of     30 Oct 82 7:34:10-EDT (Sat).
From: obrien at RAND-RELAY
Via:  Rand-Relay; 30 Oct 82 12:43-EDT

       You got it.  We had a bolus of 600 messages come in for Rand-Relay
at a time when UDel-Relay's Arpanet connection was very, very sick.
We do not have a TCP connection to UDel-relay, and the messages were
breaking their NCP, so I put them off in a hole.  When UDel was finally
fixed, I put the messages back, and they all timed out simultaneously.
The multiple instances of "deliver" pushed the load average up to 7.0
on Rand-Relay, and caused multiple outgoing copies of the returned mail.

       We have set up procedures now which will let us take over directly
the functions of a broken relay, so this sort of thing shouldn't happen
again.

----- End of forwarded messages