[19367] in Hotline Meeting

home help back first fref pref prev next nref lref last post

Steps/Process for a Patch release

daemon@ATHENA.MIT.EDU (Tim McGovern)
Thu Oct 28 11:31:38 1993

Date: Thu, 28 Oct 93 11:31:04 EST
From: tjm@MIT.EDU (Tim McGovern)
To: release-76@MIT.EDU, athena-outage@MIT.EDU, op@MIT.EDU, rel-eng@MIT.EDU
Cc: acmg@MIT.EDU, dcns-dev@MIT.EDU

Some recent confusion surrounding a Solaris patch release led Kim and me to 
review how we're coordinating such releases.  This note serves simply to remind 
everyone of the steps that we've generally been following.  We also suggest a 
few details that would lend a little more consistency in how we communicate 
where we are in the patch/change process.  We also hope that this note will be 
helpful to folks who would like to keep better informed, although they might be 
on the periphery of the release process itself.

The scope of this suggestion is where a critical bug or operational problem 
exists in the Athena environment.  It does not cover situations where real-time 
correction is necessary.

0.  rel-eng learns of a critical bug from somewhere...and they confirm
    the problem.

1.  rel-eng sends mail to the current release mailing list, 
    <release-76@mit.edu> for this release, when they begin to
    investigate a critical bug.

2.  rel-eng sends mail to <release-nn@mit.edu> when they believe 
    they have the fix in hand, and are ready to begin *any* sort of 
    testing. This includes testing from the rel-eng cell or via a fix 
    installed native on the client.

    As testing becomes more widespread, rel-eng or ops sends mail to
    a wider audience, such as the <athena-outage@mit.edu> mailing list.

    There is often going to be iteration at this stage while testing
    goes on.  At some point, rel-eng and ops will generally agree that 
    a releaseable fix has been found.

3.  rel-eng or ops sends mail to <athena-outage@mit.edu> to inform 
    everyone that a fix has been found, and deployment is imminent, 
    although a schedule may not have been set as yet.  At this stage,
    all of the affected groups need to be bought in to the deployment.

4.  ops sends mail to <athena-outage@mit.edu> indicating that
    a fix will be deployed, and the schedule.

5.  ops sends mail to <athena-outage@mit.edu> when the deployment
    has been completed.

We welcome your comments, suggestions and questions.

Thanks,
Tim

home help back first fref pref prev next nref lref last post