[Olpc-sysadmin] wiki.laptop.org upgrade

Ed McNierney ed at laptop.org
Thu Dec 4 06:25:29 EST 2008


Bernie -

We should *never* take our public-facing Web services offline deliberately
without scheduling that event in advance and sending warnings and reminders
of when that scheduled maintenance will occur.  Never.  There are a lot of
people who have been working very hard over the last few weeks to ensure
these primary Web services are available, online, and reliable - with
fallback systems in place in case something (like a software upgrade) goes
wrong.

This should never happen again.  We cannot be taking our servers offline
during a busy period of the day (late afternoon or early evening, prime time
for our major US G1G1 market) for 45 minutes while we figure out "weird
problems".

Perhaps I missed it, but I do not recall any email review or discussion of
the value or need for either OpenID or a MediaWiki update, and I don't
understand how we made the decision that either was more valuable than
keeping one of our two major public sites online.

Having volunteer assistance for systems administration is extremely
valuable, but that assistance must be coordinated and communicated with the
rest of the team.

    - Ed


On 12/4/08 5:08 AM, "Bernie Innocenti" <bernie at codewiz.org> wrote:

> Ed McNierney wrote:
>> What was the motivation for this upgrade?  Why did we need to take the wiki
>> offline for several hours during our G1G1 promotion?  Thanks.
> 
> It was offline for approximately 45 minutes (and it was mostly due to
> a weird problem that took a while to figure out).
> 
> The main motivation for the upgrade was installing OpenID to enable
> single-sign-on across all the web applications.  Secondarily, it's
> always safer to keep web applications up to date.  I also did a few
> cleanups to ensure the next updates will be a little easier.




More information about the Olpc-sysadmin mailing list