|
|
|
Communication:
|
|
|
|
There should be something on:
|
|
* news.opensuse.org (article with some details - or a link to the wiki page with details?)
|
|
* Email to opensuse-announce@opensuse.org
|
|
* Email to opensuse-project and opensuse-factory - and opensuse-buildservice
|
|
* Social media
|
|
* #opensuse-admin topic
|
|
* obs announcements feed
|
|
* forums.o.o
|
|
* maybe also https://progress.opensuse.org/projects/opensuse-admin/news
|
|
|
|
When?
|
|
* send the mail on opensuse-announce both on 2017-09-29 and a reminder 3 and 1 day before
|
|
* news.opensuse.org will see an article at 2017-09-29
|
|
* add a shortened version on progres.opensuse.org around 2017-10-06 (later as the news.o.o article, to have this news as 2nd reminder on planet.o.o)
|
|
* 2017-10-06 headers of important services like wiki, forums and OBS should be changed to announce the downtime
|
|
* 2017-10-10 Email announcements go out to:
|
|
* opensuse-project
|
|
* opensuse-factory
|
|
* opensuse-buildservice
|
|
* 2017-10-13
|
|
* change the topic headers on all #opensuse-* IRC channels
|
|
* again an announcement mail on opensuse-announce
|
|
|
|
Action items:
|
|
* cboltz, ada_lovelace: ask darix to change the channel topics on all #opensuse-* IRC channels for the downtime
|
|
* tampakrap: prepare announcements and sent them around as planned
|
|
* tampakrap: be around on IRC during the downtime, to help our customers ;-)
|
|
|
|
Services Migration:
|
|
* status2: take a recent dump from status.o.o and change dns (AI: lrupp)
|
|
* download.opensuse.org is a must have - this includes a redundant PostgreSQL cluster and a mirror machine (no scanner) (AI: lrupp)
|
|
* static.opensuse.org as well
|
|
* conncheck is also important
|
|
* freeipa.infra.opensuse.org needs a master-master setup anyway (and is needed to not loose the DNS)
|
|
* chip is also needed (the machine that transfers the dns data to MF) - not really, as the DNS servers from MF transfer only from a specific IP, which can not be running in the Provo DC. Instead, we need to enlarge the TTL for the zone transfer to at least 4 days BEFORE the downtime
|
|
|
|
The haproxy in Provo should deliver 503 pages for all other services that are not migrated (and therefor down) during the downtime.
|
|
|
|
If time permits:
|
|
* scar - to get a log in point in Provo
|
|
* Galera cluster (as slave with a bigger lag)
|
|
* lists
|
|
* wiki machines (main + elastic search) - in read only mode
|
|
* keyserver
|
|
|