[111] in Athena_Backup_System
Minutes of meeting 6/15/95
daemon@ATHENA.MIT.EDU (Jonathon Weiss)
Thu Jun 15 16:26:41 1995
Date: Thu, 15 Jun 95 16:19:06 EST
From: jweiss@MIT.EDU (Jonathon Weiss)
To: athena-backup@MIT.EDU
Here are the minutes as I electronically scribbled them down during the
meeting.
Jonathon
Attending: jweiss delgado vrt bwmelans dkk
We are going to get are own cell. It will have a real/consistant
version of AFS (ie, something that we can reproduce) (Probably what
everis running in the Athena cell, we need to have source that builds
to the binary) It will be a one DB server cell on a maxine. Toto will
be moved to the backup cell. This is for security reacons, and so that
we can have better control.
Diane mumbled something about cross realm being an issue. Diane asks
who will have access to the machine. it looks like it will be ops and
abs.
Will there be issues with network badwidth? several people don't think
this will be an issue, unless there are bugs, at least until we start
doing tape dumps.
plover may move subnets (it's on 18.72 now) also if you have a better
name for it, it may change names.
TESTING
vrt hands out a status report for what tests have been
designed/coded/run
eventually we'll need to simulate a slave so vrt can test the other side
of the master. We may need a machine for this
vrt hands out the header file with the macros that I couldn't find last
week.
Do we want the master to consider file mark size, so the master can
guess how full a tape will be. Ask the slave? How would it know? The
file mark siz will vary with the drive type, and possibly a drive type
will have more than one size. resolution: Store this info in the media
pool configurable with a default. (A media pool is defined to have one
type of tape)
Diane: we should probably have a code review after we do at least some
preliminary testing on most of the master. vrt: 100 to 200 lines / hour
up to 2 hours per day.
What about our coding style guide? the gnu guide has a lot of
irrellivant baggage. We do need a consistant way of doing some things
like braces. vrt handed out a checklist a while ago, we're going to
look at that instead.
tape slave not going as fast as jweiss would like. Did get some info
from Diane, that involved changes ti a struct or two. integrated
testing by the end of the month? possible, but not likely.
we didn't get a UROP for the GUI, we may have a lead, but we're not
sure.
we do have people for testing; cat, kpimsa, and someone else. They
should start comming to meetings, but mbarker and vrt have been having a
little trouble coordinating with them and each other.
Diane will give us some code to look at for two weeks from now. to see
if her style, etc. is what we expect.
