[1605] in SIPB-AFS-requests
Re: Cutover plans
daemon@ATHENA.MIT.EDU (mhbraun@MIT.EDU)
Tue Nov 15 19:38:31 1994
From: mhbraun@MIT.EDU
Date: Tue, 15 Nov 1994 19:35:45 -0500
To: Derek Atkins <warlord@MIT.EDU>
Cc: Greg Hudson <ghudson@MIT.EDU>, Derek Atkins <warlord@MIT.EDU>,
sipb-afsreq@MIT.EDU, yandros@MIT.EDU
In-Reply-To: [1602]
If we move both machines at once it is better the way you suggested we do it
(which is exactly how I did ronald-ann to a decstation from an RT) but it has
several drawbacks
1) outages....Greg's way has no user visible outage and if we have the swap
space it probably is the best.
2) Delni space....it is true that we can chain delnis'''but we may not *want*
to
3) Clients *will* lose if we change the address of bot hservers at one time.
jweiss and I can fix the athena Celsrvdb but clients will not see that for a
little time. We might want to move only one and tell transarc that both moved
(since they should do the dns resolve anyway) and then move the second one
later (keeping the Athena Cellsrvdb in sync with reality).
4) Doing everything at once changes many varaibles. It is not clear we want
to change that many.
This is not to say that Derek's way would not work...but if we have swap disk
then it might not be the best.
Matt
