Switch Datacenter

Introduction

A datacenter switchover (from eqiad to codfw, or vice-versa) comprises switching over multiple different components, some of which can happen independently and many of which need to happen in lockstep. This page documents all the steps needed to switch over from a master datacenter to another one, broken up by component.

Schedule for Q3 FY2015-2016 rollout

• Deployment server: Wednesday, January 20th
• Media storage/Swift: Thursday, March 10th 17:00 UTC
• Traffic: Thursday, March 10th
• MediaWiki 5-minute read-only test: Tuesday, March 15th 07:00 UTC
• Services: Thursday, March 17th, 10:00 UTC
• Services (second test): (week of March 28th)
• ElasticSearch: Thursday, April 7th, 12:00 UTC
• MediaWiki: Tuesday, April 19th, 14:00 UTC / 07:00 PDT / 16:00 CEST (requires read-only mode)

Switching back

• MediaWiki: Thursday, April 21st, 14:00 UTC / 07:00 PDT / 16:00 CEST (requires read-only mode)
• Services, ElasticSearch, Traffic, Swift, Deployment server: Thursday, April 21st, after the above is done

Per-service switchover instructions

MediaWiki-related

Before switch over (after any local testing within codfw):

• Wipe memcached to prevent stale values (MediaWiki isn't ready for Multi-DC yet).
  • Once eqiad is read-only and cofdw read-only master/slaves are caught up; any subsequent requests to codfw that set memcached are fine.

The overall plan for an eqiad->codfw switchover is:

• Warmup codfw databases
• Warming up memcached with parsercache entries
• Stop jobqueues in eqiad
• Deploy mediawiki-config with all shards set to read-only
• Set eqiad databases (masters) in read-only mode; stop pt-heartbeat (?)
• Set $app_routes['mediawiki'] = 'codfw' in puppet
• Run puppet on all redis hosts. Replication is now flowing codfw => eqiad
• Deploy Varnish to switch backend to appserver.svc.codfw.wmnet/api.svc.codfw.wmnet
• Deploy services to change their MediaWiki action API endpoint
  • RESTBase (uses puppet's $app_routes, needs puppet run + service restart)
  • Parsoid (manages its own config, will need a deploy + restart of its own)
  • All other services will be automatically reconfigured whenever puppet runs after $app_routes is modified
• Master swap for every core (s1-7), ES (es1-3), parsercache (pc) and External store (x1) database
  • Technically there is nothing to do at database level once circular replication is setup
  • In reality, some small deploys: set codfw masters mysql as read-write and start pt-heartbeat-wikimedia there; change *-master dns to the new masters (only used by humans); optionally: puppetize $master = true
• Deploy mediawiki-config (only codfw?) with all shards set to read-write
• Start the jobqueue in codfw

The plan for switching back is the reverse of the above, with the following extra step:

• Wipe memcached to clear invalidated cached memory

Databases

See the separate page on how to promote a new slave to master. Please note that no topology change happens on failover, so very little from that applies.

Job queue

• Jobrunners in eqiad get stopped. This is done by setting jobrunner_state: 'stopped' in hiera
• mediawiki goes read-only - this should ensure no new job gets enqueued.
• $mw_primary is set to 'codfw
• mediawiki primary gets switched in mediawiki-config
• force a puppet run on the codfw redis hosts and on eqiad hosts after that: salt 'rdb2*' cmd.run 'puppet agent --enable; puppet agent -t'
• mediawiki is set read-write in codfw
• We start the jobrunners in codfw and they will consume the jobs left over from eqiad. This is done by setting jobrunner_state: running

Debugging

You can force Varnish to pass a request to a backend in codfw or eqiad using the X-Wikimedia-Debug header.

For codfw, use X-Wikimedia-Debug: backend=mw2017.codfw.wmnet

For eqiad, use X-Wikimedia-Debug: backend=mw1017.eqiad.wmnet

Media storage/Swift

Originals

• Instruct mediawiki to write synchronously to both eqiad/codfw with https://gerrit.wikimedia.org/r/276071
• Change varnish backends for app swift in hieradata/common/cache/upload.yaml to point to codfw
• Change route_table in hieradata/role/common/cache/upload.yaml so that codfw is direct.
• Change route_table in hieradata/role/common/cache/upload.yaml so that eqiad points to codfw.

Thumbnails

• Point swift rewrite middleware to codfw with https://gerrit.wikimedia.org/r/#/c/268080/ and run puppet on ms-fe* plus swift-init all reload for the changes to take effect. Now 404s for thumbs will hit codfw imagescalers.
• Change varnish backends for app rendering in hieradata/common/cache/text.yaml to point to codfw
• Change varnish backends for app swift_thumbs in hieradata/common/cache/upload.yaml to point to codfw, user traffic for thumbs hits swift codfw

ElasticSearch

Point CirrusSearch to codfw by editing wmgCirrusSearchDefaultCluster InitialiseSettings.php. The default value is "local", which means that if mediawiki switches DC, everything should be automatic.

Traffic

GeoDNS user routing

• Traffic-layer only, no interdependencies elsewhere
• Granularity is per-cache-cluster (misc, maps, text, upload)
• Documented at: https://wikitech.wikimedia.org/wiki/Global_Traffic_Routing#GeoDNS

Inter-Cache routing

• Traffic-layer only, no interdependencies elsewhere
• Granularity is per-cache-cluster (misc, maps, text, upload)
• Documented at: https://wikitech.wikimedia.org/wiki/Global_Traffic_Routing#Inter-Cache_Routing

Cache->App routing

• Normally will have inter-dependencies with application-level work
• Granularity is per-application-service (how they're defined at the back end of varnish)
• Documented at: https://wikitech.wikimedia.org/wiki/Global_Traffic_Routing#Cache-to-Application_Routing

Services

• RESTBase and Parsoid already active in codfw, using eqiad MW API.
• Shift traffic to codfw:
  • Public traffic: Update Varnish backend config.
  • Update RESTBase and Flow configs in mediawiki-config to use codfw.
• During MW switch-over:
  • Update RESTBase and Parsoid to use MW API in codfw, either using puppet / Parsoid deploy, or DNS. See https://phabricator.wikimedia.org/T125069.

• Tracker / checklist

Other miscellaneous

• Deployment server
• EventLogging
• IRC/RCstream