incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Henri Yandell <>
Subject Re: The Incubator and Infrastructure
Date Fri, 02 Sep 2005 08:55:23 GMT
On 9/2/05, Niclas Hedhman <> wrote:
> On Friday 02 September 2005 10:28, Davanum Srinivas wrote:
> > Several people on WS-PMC offered to help infra. They were turned down
> > as they are not members.
> This has been my impression as well.
> infra@ is begging for help, but seems to have problem with;
>  a. only members allowed in for the stuff that each projects
>     PMC Chair can't do already.
>  b. requesting people without access to help out on answering
>     questions, which more often than not, requires peeks into
>     files that such volunteer doesn't have access to, or to
>     places that may or may not be documented.

As someone currently sinking into the mire of infra having just become
a member, I have to agree; the work available for the average
committer isn't very clearly identified.

This has improved, Bugzilla->Jira migrations are now in danger of
happening and although Henning is a member, Tim isn't and I don't
think Henning's member-ness is relevant to the work.

Answering user questions on the infra@ mailing list is always useful,
especially given the infra-private@ bit below.

> Leo,
>   There are three ways to ease the burden on infra team, and I think all are
> wanted;
>    1. reduction of requests coming into infra@ (this thread covers that),

Also partly solved in a different way. infra-private was created for
dealing with the emergencies/big-bad-things (hard-drive blown up,
minotaur fscks, giving access to volunteer xxxx).

The infra@ mailing list has moved more towards user requests and some
of the core people have the chance to opt out of this list now.

>    2. more hands to help out,

Thus the volunteer requests.

>    3. more tools to automate.

Leo's been behind a lot of design plans for that, and there's a CA
tool looming on the horizon. An improvement here would be for someone
(Leo?) to give us a better overall plan of what's requested and where
it fits.

One of the problems with this is that you still need a certain access
to handle some of them, or at least a certain amount of knowledge of
how the system works.

I think there may be an infra-tools@ list that I didn't join to
discuss these. Can't remember if that happened or not.

There's another one too. Get involved with the site-dev@apache mailing
list (I think that's the correct one?); discussing ways to provide a
site building zone and how to handle that kind of stuff.

> I am specially curious of what can be done about 3.
> ASF consists of (allegedly) more than a 1000 committers. How hard could it be
> to find a couple that are capable of putting this in place? Very. Since only
> members are allowed in, they are busy elsewhere and infra team has no time to
> extract all the details for someone to implement these independently, or so
> it seems.

We're all here to scratch our itches, writing code for Infra doesn't
scratch the itch unless you've been sucked into Infra. But you're
right, cornering Leo into drawing a picture of Infra, what bits are
where and what bits can be automated would be useful. Leo, will you be
at ApacheCon? :)

> As for more hands, I think the infrastructure team needs to compile a list of
> what committers, members and root must do. Once you have identified work that
> "committers" can do, members should volunteer higher up the ladder, and
> likewise for roots.

Yup. My first worry was on what the existing people were doing and
where we were lacking backup, so I've a very rudimentary form of this
with people's actual names attached. I'll work to genericize this and
try to put in more tasks, and hopefully find out where there are
actual problems. SVN migrations are ticking along quite nicely for
example, so no great need to throw more resource at it (unless it's
people telling us that they're ready to migrate, hint hint).


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message