Scrum of scrums/2013-11-19
Docs
[edit]http://etherpad.wikimedia.org/p/Scrum-of-Scrums
Notable action items
[edit]- W0 awaiting ops approval on X-Forwarded-By header work - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/38
- Mobile Web still blocked by GettingStarted API from Growth team - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/11
- Mobile Web affected by ULS bug from Language team - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/45
- Language still blocked on emulating replag in dev environment; have been corresponding with SPringle - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/30
- Flow has ongoing architecture review form ops; awaiting more feedback - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/46
W0
[edit]Adam Basso
- (Ops approval + Analytics follower) X-Forwarded-By header work with Faidon (Ops) - Faidon, Yuri, Adam to figure out short/med/long term options
- (Ops implement vhost portion + Analytics follower) Landing page for m.wikipedia.org and zero.wikipedia.org with Apache vhosts coming up in the next (hopefully) 1-2 weeks- Yuri to work with Faidon
- (Mobile Apps approval and knowledge transfer) Firefox OS Wikipedia app bugfixes and deploy - needs to be done by tomorrow to make quarterly cutoff for inclusion in base OS. Yuri likely person to deploy in mediawiki-config tomorrow pending today's review.. Adam to talk with Brion.
Parsoid
[edit]Gabriel Wicke
- Had quarterly planning meeting
- Deployed twice: https://www.mediawiki.org/wiki/Parsoid/Deployments
- Public & cached parsoid service at http://parsoid-lb.eqiad.wikimedia.org/, implementing https://www.mediawiki.org/wiki/Parsoid#The_Parsoid_web_API
- Platform still blocked on article transfer to betalabs by ops, pending available resources - https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/37
Ops
[edit]Faidon Liambotis and Andrew Otto
- Varnish serving all traffic since Nov 11th; a few issues found & fixed
- Swift migrated to eqiad (last part of production traffic)
- An outage a day this past week - see ops@ list for post-mortems (soon on wikitech, hopefully)
- ulsfo deployment paused, back to procurement to workaround network/vendor issues
- Assisting in the PDF sprint (see separate update by Matt)
- aka "collectoid" or "ocg" (offline content generation)
- Labs
- New contractor (Mike Hoover) hired to assist with migrating to eqiad
- Analytics nodes reshuffled into multiple rows for redundancy, still a few issues to deal with:
- analytics1013 and analytics1021 seem to have the same IP assigned, Chris J is looking into it.
- Need publicly routable IPv6 for analytics1021 and analytics1022 Kafka Brokers.
- Lots of RFPs for Datacenters being reviewed.
- New search servers racked, working with Nik to get Elasticsearch ready on them.
- Search outage caused by chain of bad assumptions, lsearchd (Lucene) currently turned back on for all wikis until issues are resolved.
- Card 30 -- unblocked
- Card 37 -- still blocked on resources
- Card 19 -- done
PDF Handler
[edit]Matt Walker
- Goal is deploying to beta cluster this week; will need another week to deploy
Growth
[edit]Matt Flaschen
- Progress on keep going (refactoring); API module next
- Planning for next project -- supporting article creation workflow
Mobile Apps
[edit]Brion Vibber
- Firefox OS Wikipedia App work (W0 taking responsibility, have high priority bugs to fix this week to make quarterly cutoff)
Language
[edit]Niklas Laxstrom
- Team at conference
Analytics
[edit]Dan Andreescu
- Geowiki info leak resolved
- Ongoing -- migrating Wikimetrics to production (ops)
- Cards are in progress
Mobile
[edit]Jon Robson
- Team at language summit; mostly bug fixing
- Exposed bug with ULS (Language)
- Working on Limn issue with Analytics (Analytics)
Core features (Flow)
[edit]S Page
- mediawiki.org deploy early December
- Parsoid intergration going well
- Ops review needed -- pass bug to Faidon(Ops)