[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Administrivia



SRI was unable to talk to portions of the MILNET (and hosts behind that, from
the NSFNET view) from Thursday morning until Monday night.  The problem was not
within SRI but apparently was at the NSFNET/MILNET boundary.  Those people on
SLUG that are on Milnet hosts may have lost some mail. (I show 12 SLUG messages
with a date of 22-Mar through 25-Mar mainly discussing "Bundling of layered
software".)

The hosts I had mail queued up tonight before things got fixed were:
bbn.com, brl.mil, etl.army.mil, tycho.ncsc.mil, HUACHUCA-EMH2.ARMY.MIL,
rtc.reston.unisys.com, alexandria-emh2.army.mil, dockmaster.ncsc.mil,
westpoint-emh1.army.mil, sandia.gov, nprdc.navy.mil,
aio.jesnet.jsc.nasa.gov, PAXRV-NES.NAVY.MIL,
sccgate.scc.com, NSFnet-Relay.AC.UK, aagate.avlab.wpafb.af.mil,
cs.sandia.gov, tops20.radc.af.mil

If any of the people on those hosts want the mail resent to them, please
let SLUG-Request@AI.SRI.COM know.

This also explains why mail from those hosts to SLUG (and SRI in
general) was not getting through.


Background:  Two major changes happened recently.  The one that
apparently is related to this is that NIC.DDN.MIL (formerly
SRI-NIC.ARPA) used to be hooked up to ARPANET and MILNET IMPs here.  The
ARPANET IMP recently went away.  However, there must have some routes
that were specialized to know that SRI had its own route to the ARPANET
and MILNET.  When our IMP went away, these routes had to be removed.
(Actually, what was happening was that gateways were too smart and
refused to advertise our routes, first on BARRNET and then on MILNET.)

The second change that happened was that we reconfigured our network in
the AI Center here.  We used to have our own network with one gateway to
the rest of SRINET (and, more than a year ago, our own gateway to
ARPANET).  Our hosts are no directly on a subnet of SRINET.  That
happened 2 weeks ago.  There were some problems for a day or so and then
a week later the new gateway we were plugged into was down for 36 hours
over the weekend.  This, and another local lossage, caused us to change
AI.SRI.COM's network address for a few days.  (Currently it still has
its old address but also has a new address on SRINET.)  I don't think
anyone should have noticed it but if you did, that is what happened.