discourse/app/jobs/regular/group_smtp_email.rb
Martin Brennan b3d3ad250b
FIX: Handle SMTPServerBusy for group smtp email (#13632)
Use the `sidekiq_retry_in` code from Jobs::UserEmail in group SMTP. Also we don't need to keep `seconds_to_delay` -- sidekiq uses the default delay calculation if you return 0 or nil from the block. See 3330df0ee3/lib/sidekiq/job_retry.rb (L216-L234) for sidekiq default retry delay logic.

I experimented with extracting this into a concern or a module, but `sidekiq_retry_in` is quite magic and it would not allow me to abstract away into a module that calls some method specificall in the child job class.

I would love to write tests for this, but it does not seem possible (not sure if its because of our test
setup) to write tests that test sidekiq's retry capability, and I am not sure if we should be anyway. Initial addition
to UserEmail did not test this functionality 
d224966a0e
2021-07-06 13:37:52 +10:00

101 lines
3.4 KiB
Ruby

# frozen_string_literal: true
require_dependency 'email/sender'
module Jobs
class GroupSmtpEmail < ::Jobs::Base
include Skippable
sidekiq_options queue: 'critical'
sidekiq_retry_in do |count, exception|
# retry in an hour when SMTP server is busy
# or use default sidekiq retry formula. returning
# nil/0 will trigger the default sidekiq
# retry formula
#
# See https://github.com/mperham/sidekiq/blob/3330df0ee37cfd3e0cd3ef01e3e66b584b99d488/lib/sidekiq/job_retry.rb#L216-L234
case exception.wrapped
when Net::SMTPServerBusy
return 1.hour + (rand(30) * (count + 1))
end
end
def execute(args)
return if quit_email_early?
group = Group.find_by(id: args[:group_id])
return if group.blank?
post = Post.find_by(id: args[:post_id])
email = args[:email]
cc_addresses = args[:cc_emails]
recipient_user = User.find_by_email(email, primary: true)
if post.blank?
return skip(email, nil, recipient_user, :group_smtp_post_deleted)
end
if !Topic.exists?(id: post.topic_id)
return skip(email, post, recipient_user, :group_smtp_topic_deleted)
end
if !group.smtp_enabled
return skip(email, post, recipient_user, :group_smtp_disabled_for_group)
end
# There is a rare race condition causing the Imap::Sync class to create
# an incoming email and associated post/topic, which then kicks off
# the PostAlerter to notify others in the PM about a reply in the topic,
# but for the OP which is not necessary (because the person emailing the
# IMAP inbox already knows about the OP)
#
# Basically, we should never be sending this notification for the first
# post in a topic.
if post.is_first_post?
ImapSyncLog.warn("Aborting SMTP email for post #{post.id} in topic #{post.topic_id} to #{email}, the post is the OP and should not send an email.", group)
return
end
ImapSyncLog.debug("Sending SMTP email for post #{post.id} in topic #{post.topic_id} to #{email}.", group)
# The EmailLog record created by the sender will have the raw email
# stored, the group smtp ID, and any cc addresses recorded for later
# cross referencing.
message = GroupSmtpMailer.send_mail(group, email, post, cc_addresses)
Email::Sender.new(message, :group_smtp, recipient_user).send
# Create an incoming email record to avoid importing again from IMAP
# server. While this may not be technically required if IMAP is not
# currently enabled for the group, it will help a lot with the initial
# sync if it is turned on at a later date.
IncomingEmail.create!(
user_id: post.user_id,
topic_id: post.topic_id,
post_id: post.id,
raw: message.to_s,
subject: message.subject,
message_id: message.message_id,
to_addresses: message.to,
cc_addresses: message.cc,
from_address: message.from,
created_via: IncomingEmail.created_via_types[:group_smtp]
)
end
def quit_email_early?
SiteSetting.disable_emails == 'yes' || !SiteSetting.enable_smtp
end
def skip(email, post, recipient_user, reason)
create_skipped_email_log(
email_type: :group_smtp,
to_address: email,
user_id: recipient_user&.id,
post_id: post&.id,
reason_type: SkippedEmailLog.reason_types[reason]
)
end
end
end