[205] in Athena_Backup_System
ABS minutes for 2/29
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Thu Feb 29 17:08:06 1996
Date: Thu, 29 Feb 1996 17:08:36 -0500
To: athena-backup@MIT.EDU
From: jweiss@MIT.EDU (Jonathon Weiss)
Attending: jweiss wdc vrt delgado bwmelans
Quickstart status: all done.
cleanly: some mail has been sent, but we wanted to talk about it at a
meeting, and Dave isn't here. push till next week. (Some of the
concerns are that the def'n of cleanly Dave suggested implied a
deliverable of rc scripts which we don't currently have as a
deliverable; there are also concerns about potential conflicts between
integrity constrains that should prevent he master from starting
occuring during operation. vrt will make a pass at this and try to
determin what is realistic.
Dave showed up, so we'll discuss this now.before we can commit to
numbers, we want some idea of what oracle will commit to for re-runing
logs, etc. Numbers also depend on what constraints we require on
operating the DB. disk errors are obiviously a more serious than a
crash w/o fs corruption.
Is an rc script for the master a deliverable? Is it acceptable for
starting the master process to be a manual operation? yes, assuming
oracle has reasonable tools for insuring consistancy. but it's not
ideal. If we can provide tools to automate this, that would be nice
for the subsequent version.
jweiss will create a discuss meeting for the abs-alpha.
For next week (bill will be out)
possibly cripple mode stuff
review do_dump() take 3
Jonathon
