[google-cloud-datastore] timeout option has no effect: per-RPC defaults override it (same as #14050)

Ouverte
#36,632 1 commentaire 0 réactions 0 personnes assignées Voir sur GitHub

Les mainteneurs répondent en général sous 1 jour

Personne n'a encore pris cette issue.

Évaluation

Cette issue n'a pas encore été évaluée.

Description

api: datastore priority: p2 type: bug

The timeout option documented in OVERVIEW.md ("Configuring timeout") is accepted by Google::Cloud::Datastore.new but never applied to requests. Every RPC is sent with the 60 s RPC-specific default from google-cloud-datastore-v1.

This is the same problem reported for Pub/Sub in #14050 and fixed there in #14405 ("Propagate timeout to client RPC configs"). The Datastore library didn't get the equivalent change.

Environment details
  • OS: macOS
  • Ruby version: 3.4.9
  • Gem name and version: google-cloud-datastore 2.14.0, google-cloud-datastore-v1 1.6.0, gapic-common 1.3.0

The code involved is unchanged on main.

Steps to reproduce
  1. Create a dataset with timeout: 5.
  2. Make any request.
  3. The request is sent with a 60 s timeout, not 5 s.
Code example
require "google/cloud/datastore"

ENV["DATASTORE_EMULATOR_HOST"] = "localhost:8081" # no server needed; the RPC is intercepted

datastore = Google::Cloud::Datastore.new project_id: "my-project", timeout: 5
client = datastore.service.service # Google::Cloud::Datastore::V1::Datastore::Client
puts "client config.timeout:      #{client.configure.timeout.inspect}"
puts "client config.rpcs.commit:  #{client.configure.rpcs.commit.timeout.inspect}"

stub = client.instance_variable_get :@datastore_stub
stub.define_singleton_method :call_rpc do |method, _request, options:|
  puts "#{method} sent with timeout: #{options.timeout.inspect}"
  raise "stop"
end

[-> { datastore.save datastore.entity("Task", "a") },
 -> { datastore.find "Task", "a" },
 -> { datastore.allocate_ids datastore.key("Task"), 1 }].each do |call|
  call.call
rescue StandardError
  nil
end

Output:

client config.timeout:      5
client config.rpcs.commit:  60.0
commit sent with timeout: 60.0
lookup sent with timeout: 60.0
allocate_ids sent with timeout: 60.0
Cause

Google::Cloud::Datastore::Service#service sets only the client-wide config.timeout (service.rb:48). The generated client sets rpcs.<rpc>.timeout = 60.0 for every Datastore RPC and applies it first:

options.apply_defaults timeout: @config.rpcs.lookup.timeout, ...  # 60.0
options.apply_defaults timeout: @config.timeout, ...              # 5

Gapic::CallOptions#apply_defaults only fills an empty value (@timeout ||= timeout), so the client-wide timeout is never used.

Suggested fix

Apply the same change as #14405 to google-cloud-datastore: when a timeout is given, copy it to each config.rpcs.*.timeout in Service#service. Failing that, update OVERVIEW.md to say the option has no effect and document the per-RPC workaround (Google::Cloud::Datastore::V1::Datastore::Client.configure { |c| c.rpcs.commit.timeout = ... }).

Langage dominant
Ruby
Étoiles
1.4k
Forks
581
Merge moyen
1 j 22 h
PR mergées (30 j)
69

Préparer son environnement

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de googleapis/google-cloud-ruby

Toutes les issues de googleapis/google-cloud-ruby

Issues similaires

Plus d'issues Ruby

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.