UMCE linux April 6, 2004 meeting notes

[Patrick M McNeal <mcneal@xxxxxxxxx>]

April 6, 2004
Facilitated by Jane


Attending: Jane, Sean, Wes, Patrick, Andrew, Aaron, Albert, Kevin, 
Rodger, Martin, Jeff, Rick, Adi, Tony, Kevin, Marcus

Agenda:
    Next meeting notes & facilitator
        - martin will facilitate
        - Notes : Patrick

    Agenda Review
    Introductions

    Kernel Configuration (Sean, Patrick, Jane)
        - for use on server, desktop and bootcd
    	- packet filter modules selected
        - implications of not forking!
        - require notification of hardware retirement
        - two competing philosophies
            - only have minimal set of software on a given host
            - try not to fork
        - security considerations with kernel bloat
        - wes suggested only turn on options when they are actually 
needed
        - sites is willing to remove things that are not needed 
immediately
        - module loading allows items to be added as overloads
        - Does turning on a module alter the kernel in any way?
            - Reasonable to assume yes.
        - Splitting out modules adds work to the overload creation 
process
        - Martin would rather split kernel then break apart module 
transcripts
        - It's okay, and appropriate for the kernel group "to fight the 
dragon" (that exist)
        - marcus has an existing kernel that people can try
            - IDE is built as a module
            - discover is used to interrogate hardware and load modules
        - Kevin Jones pointed out that since we are small, we should be 
able to workout our issues
        - Kernel configuration group should explore module management 
further
        - There will be another kernel meeting.  e-mail sweda@xxxxxxxxx 
if you would like to attend

    Clean up of Radmind Server
        -OS group layout of the file system
            L1/ - current (first rev UMCE Linux)
            L2/ - UMCE 2.0 transcripts
            Sol8/ - solaris 8 stuff
            (grp) web
            (usr) mcneal

            enclosing top level directory tells you how stable the 
transcripts are.
        -There appears to be agreement on the structure
        -A time frame will be determined, most likely a flag day

    CD group update
        -Attempted to meet yesterday
        -working on simplifing the process of creating boot cds
        -two boot cds, one automated
        -speed issue seem to not be there
        -same toolbox on both kits
        -radmind to manage the cd iso

    OS group update
        - met week ago yesterday
        - contents lists for tentative transcripts
        - still a few things we don't quite know where to put yet
        - working on the "why" things are where they are.
        - after that we should be in a position to start building 2.0 
stuff.

    sites update
        - sites is doing great
        - X plus a bunch of apps are working
        - need authentication pieces (pam) for deployment
        - 80% to deployment goal of cyber terminals (phase 1)

    tardiness (He, he)
        - so do we need a quorum call? Michigan time?
        - First week of the month we get pushed into this crappy room
        - Kevin Jones noted that if we plan on starting on Michigan 
time, there is possibility that meeting will start even later.
        - To motivate people to show up on time, it was suggested that 
late folks may add agenda items to the end of the agenda. And that 
these late additions to the agenda will only be discussed if there is 
time left. Otherwise they will be tabled for a later meeting.
        -so the consensus is to start at 3 as much as possible, tardies 
may add agenda items to the end of the agenda, but if we run out of 
time your SOL.