WIP: Cluster "Search Servers" config and back-end abstraction
AbandonedPublic
Actions

Authored by 20after4 on Feb 24 2017, 9:29 AM.

Details

Reviewers

None

Group Reviewers

Blessed Reviewers

Summary

The goal is to make fulltext search back-ends more extensible, configurable and robust.

When this is finished it will be possible to have multiple search storage back-ends and
potentially multiple instances of each. Individual instances can be configured as any of:

read-only
write-only
read-write
disabled

I'm also attempting to reuse the connection health monitoring infrastructure from
PhabricatorDatabaseHealthRecord for monitoring the health of elasticsearch nodes. This will
allow Wikimedia's phabricator to be redundant across data centers (mysql already is,
elasticsearch should be as well).

The real-world use-case I have in mind here is writing to two indexes (two elasticsearch clusters
in different data centers) but reading from only one. Then toggling the readable property when
we want to migrate to the other data center (and when we migrate from elasticsearch 2.x to 5.x)

Remaining TODO:

Callers will obtain a wrapper that will have the PhabricatorFulltextStorageEngine interface but will

transparently aggregate multiple back-end instances.

Test Plan

Still work in progress.

Diff Detail

Repository

rP Phabricator

Branch

master

Lint

Lint Passed

Unit

Tests Passed

Build Status

Buildable 15751
Build 20817: Run Core Tests
Build 20816: arc lint + arc unit

Event Timeline

20after4 created this revision.Feb 24 2017, 9:29 AM

Herald added a reviewer: Blessed Reviewers. · View Herald TranscriptFeb 24 2017, 9:29 AM

Herald added a subscriber: epriestley. · View Herald Transcript

Harbormaster completed remote builds in B15751: Diff 41854.Feb 24 2017, 9:31 AM

As currently implemented, search.servers looks like this:

"search.servers": [
  {
    "type": "elasticsearch",
    "hosts": [
      {
        "host": "localhost",
        "roles": [ "read", "write" ]
      }
    ],
    "port": 9200,
    "protocol": "http",
    "path": "/phabricator",
    "version": 5
  },
  {
    "type": "mysql",
    "roles": [ "write" ]
   }
],

Settings at the 'host' level can override the settings from one level higher, so, for example, the individual host entries can specify port or version to override the service-level definition.