[news.sysadmin] att

jhc@att.ATT.COM (Jonathan Hawbrook-Clark) (09/25/88)

The 'att' site in Middletown NJ lost its /usr/spool
file system over Friday night. Damage was severe but
not terminal, I'm fairly sure that a few pieces of
mail vanished, so if you sent any mail through here
in that timeframe then you might want to resend it.
-- 
Jonathan Clark		jonathan@mtune.att.com, attmail!jonathan
Any affiliation is given for identification purposes only.

The Englishman never enjoys himself except for some noble purpose.

les@chinet.UUCP (Leslie Mikesell) (09/29/88)

>The 'att' site in Middletown NJ lost its /usr/spool
>file system over Friday night. 

Not to pick on this particular incident, but this seems like a general
problem.  Has anyone implemented a scheme to automatically notice
when mail is lost?  I work with a proprietary message system (actually
I ported it to unix keeping parts of an old design) that expects
tty-like devices (i.e. unreliable) as the end points.  Each message
is assigned an input sequence number and when it is sent out, it 
gets a time-stamp and an output sequence number added to the bottom.
Each morning, a message is sent to each station with the number of
messages it sent and (should have) received the previous day.  Thus
if the operator logs the sequence numbers she will notice any missed
messages and can recall them.  Obviously, this is made possible by
running everything through a single hub, but I would like to know if
any similar multi-host systems exist.  I am starting to gateway into
other delivery methods and would like to keep the reliability that
we have now.

Les Mikesell