[lopsa-discuss] Interruptions coverage...

[Tom Perrine]

Benjamin Feen writes:
 > Time Management for System Administrators describes a pretty common practice:
 > having a rotation of people fielding interruptions for the team.
 > 
 > How does your site handle this?  What hard lessons or subtle points have been
 > addressed as your process matures?

I independently "solved" this in my department earlier this year,
shortly after we had 5 sysadmins.  Now we have 7, and it's still
working, although not everyone takes a turn (yet).

Each day there is a "sysadmin in the NOC", who answers all incoming
trouble tickets, random phone calls, and emails to "sysadmin" (a
legacy email address).  They are acting as what Tom L calls the
"mutual interruption shield" for the rest of the department.  They are
expected to handle almost all interruptions, but will sometimes have to
escalate to someone else for specialized stuff.

In exchange for a day of interruptions, they get 4 days of project
time.

I'll be re-evaluating this rotation schedule, as we now have 7
sysadmins (plus 2 security and one web person), and will have 9
sysadmins by mid-Jan.  I'm hoping to get to each person being in the
NOC once each two weeks, would give LOTS of project time.

The Networking group (separate) liked what they saw, and now they have
a network admin in the NOC each day, too.

Issues (all solvable):

1. The newest and/or most junior people escalate more than they'd like,
but they are learning a lot and catching on fast.

2. The admin who is primarily Windows is not yet fully up to speed on
Linux and can't handle some of the weirder stuff with some of the more
esoteric automated tools (yet).

3. Some of the Linux admins (especially the most junior) aren't fully
up to speed on Windows AD administration, etc.

4. This is not integrated with any after-hours on-call plan, but
probably should be.

We'll address all this in the new year, especially as soon as the two
new admins start.  I've been holding off re-arranging anything until
they show up....

In other words:

It works.  It makes people happier (most of the time).  You get more
project time for people.  Training is still very important, maybe even
more important.  Plan on even more cross-training, if you hadn't
already. Integrate this with or coordinate with other support groups.

--tep


-- 
Tom Perrine - tperrine at scea.com
Sony Computer Entertainment America