Scrum of scrums/2013-10-29
Jump to navigation Jump to search
Notable Action items
- Ops to speak to Zero team (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/17)
- Growth to contact Platform RE redis bugs (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/21)
- Lang/ops to sync up about replag testing environment (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/30)
- Apps to speak to platform/ops RE googlebot (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/18)
- Mobile web needs support from QA RE getting jenkins running qunit tests (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/31)
- Flow to coordinate with Ops/GregG RE early deploy to mw.o (https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/24)
- Working on the new main page for translatewiki.net.
- Working on stats for Google on translations in Wikimedia and translatewiki.net. In contact with Dan for that. Will need to run some scripts on the cluster.
- Translate is running into multiple higher severity issues that are caused by master/slave setups in Wikimedia and the inherent replication lag it sometimes suffers form, however small. There is no development environment where this can be tested. How can we get such an environment, so that we can fix the open issues, and in the future possibly even create tests that count on the replication lag? IMO the ideal situation would be that beta labs always has at least one second of replication lag, and that there's a puppet package that allows setting up a master/slave lagged environment in labs.
- Faidon suggests to get in contact with Ops.
- New data center RFP. No updates. Lots of emails.
- In ulsfo everything is up except text varnishes. MarkB to work on that before he;s on vacation mid november.
- LVS Ip changes for WP Zero started in ulsfo.
- Kafka is looking good. Was worried, but looks like redelivering works as designed. Possible duplicates need to be removed in hdfs before analysis.
- OSM Packages are being built and deployed later.
- varnishkafka on track for deployment on mobile varnishes.
- Want to deploy varnishkafka everywhere and turn off varnishncsa. Need to modify udp2log to consume from kafka to be backwards compatible.
- Card on the wall with dep on WP Zero: Faidon reports that Adam is to talk abotu this wit hthe Zero team. https://mingle.corp.wikimedia.org/projects/scrum_of_scrums/cards/17
- Want to make a change to robots.txt for Google bot. Will coordinate with platform or ops later this week.
- regression bugs in VisualEditor taking time
- bringing browser tests for language features up to speed, both regression tests and new features
- new QA staff for VisualEditor coming on board
- may need to spend significant time on UploadWizard soon
- Search is blocked on available hardware in equiad. Not immediate block, but critical.
- MarkB indicated there is some capacity to continue.
- Greg need more hardware for full rollout.
- Minimised loaded JS
- Put code on deployment train (A/B test from SoS wall)
- Added cookcie to mobile editors.
- No changes to wall. All just started working on those.
- Planning: Enable red links on beta in mobile (page creation on mobile)
- Matt: no status on <something>. High on the list, but not ready.
- Guided tour fixes
- Future changes to mediawiki.ui
- REST tryout planned.
- Some REDIS bugs that we need help from platform with. Not urgen.
- (Sorry about the audio, $#@! Ubuntu sound control panel only worked _after_ I quit the hangout)
- Card #15: Parsoid team is working on the API Flow needs
- Card #22/#23: VE team split out oo-ui.js, I started documenting it for ErikM cross-team UI consistency project
- NEW card #24: we'd like to eat our own dogfood by deploying Flow to mediawiki.org activated on just two talk pages, we'll send e-mail to Ops and talk to Greg G
- Card #2: Yuri is debugging ESI
- Card #3: Mark provided an IP address migration plan to support HTTPS and all languages on all projects. Schedule of implementation pending on Mark.
- Card #3 (incidental): W0 is starting to hide UI chrome for HTTPS until IP address-based zero-rating is supported. Implementation in next 9 days.
- Card #16: FlaggedRevs will work for stage-gated carrier configuration changes by carriers. Adam and Yuri to work on technical approach.
- Card #17 Varnish configs: mailing to Faidon, will have videoconference to eliminate confusion. Will talk with Diederik on Thursday to resolve Analytics dependency.
- Card #18 on googlebot can be removed as a W0 dependency
- Cache vary header: now varied on source proxy as well, so that banners are only shown if proxy zero-rating is supported at a carrier for a given proxy.
- Cassandra / Rashomon testing going well
- Temporary public API next days (for mobile, Flow, Kiwix), to be followed by proper REST content API