[Labs-l] Reminder: OpenStack Nova upgrade scheduled for Monday, August 20
Ryan Lane
rlane at wikimedia.org
Thu Aug 23 23:23:42 UTC 2012
This is still planned to go forward on Monday. There's a possibility
of extended downtime during a number of the parts of the process.
We'll try to minimize downtime, though.
On Fri, Aug 17, 2012 at 5:32 PM, Ryan Lane <rlane32 at gmail.com> wrote:
> So, an update: we'll be pushing this back a week to Monday Aug 27th.
> We've made good progress, but testing isn't totally complete. Rolling
> back the upgrade will be very difficult, so we want to do a decent
> amount of testing. So far we've completed the following:
>
> 1. Updated puppet configs to support diablo and essex, and multiple regions
> 2. Updated OpenStackManager to support:
> * Nova API
> * Essex release of OpenStack
> * Multi-regions
> 3. Brought up the eqiad region to test Essex
> 4. Installed MediaWiki with new OpenStackManager in eqiad for testing
>
> What still needs to be done:
>
> 1. Full testing of OpenStackManager against a production like environment
> 2. Full infrastructure testing of OpenStack Essex in a production like
> environment
>
> - Ryan
>
> On Mon, Aug 13, 2012 at 12:58 PM, Ryan Lane <rlane at wikimedia.org> wrote:
>> We're planning on upgrading OpenStack Nova on Monday August 20th.
>> We'll try to keep downtime to a minimum, but it's possible that there
>> will be extended downtime during the upgrade. Our deployment window
>> will be from 10:00 - 18:00 PDT. The schedule for this is ambitious, as
>> we may not be ready to upgrade by this date. If we are not ready,
>> we'll push the window to the following week (August 27th, same time).
>>
>> During this deployment window, we'll try to keep instances
>> operational, however, all OpenStack services will be down. This means
>> labsconsole access will be limited, and many instance actions won't be
>> available.
>>
>> Please do *not* reboot instances during this deployment window. Since
>> the nova-network service will be down, it's possible that rebooting an
>> instance will cause it to be down until the entire upgrade is
>> complete.
>>
>> - Ryan
>>
>> _______________________________________________
>> Labs-l mailing list
>> Labs-l at lists.wikimedia.org
>> https://lists.wikimedia.org/mailman/listinfo/labs-l
>
> _______________________________________________
> Labs-l mailing list
> Labs-l at lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/labs-l
More information about the Labs-l
mailing list