[GitHub] [couchdb] jaydoane commented on a diff in pull request #4284: Remove all usage of global

[GitBox <gi...@apache.org>]

jaydoane commented on code in PR #4284:
URL: https://github.com/apache/couchdb/pull/4284#discussion_r1037547724


##########
src/couch_replicator/src/couch_replicator_pg.erl:
##########
@@ -0,0 +1,86 @@
+% Licensed under the Apache License, Version 2.0 (the "License"); you may not
+% use this file except in compliance with the License. You may obtain a copy of
+% the License at
+%
+%   http://www.apache.org/licenses/LICENSE-2.0
+%
+% Unless required by applicable law or agreed to in writing, software
+% distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+% WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+% License for the specific language governing permissions and limitations under
+% the License.
+
+% Use pg process groups to reduce the chance of duplicate replication jobs
+% running on the same cluster.
+%
+% A custom replicator pg group is started via start_link/0. Then, replication
+% jobs check if they would be leaders before starting. If, by chance, two jobs
+% with the same RepId start anyway, then replication jobs would do an extra
+% check before each checkpoint. If the they are not leaders any longer, they
+% should stop running. The "leader" is just the first sorted element in the
+% [node(Pid), ...] list.
+
+-module(couch_replicator_pg).
+
+-export([
+    start_link/0,
+    join/2,
+    leave/2,
+    pids/1,
+    should_start/2,
+    should_run/2
+]).
+
+% Start a custom pg group. Should be called from the replication supervisor.
+%
+start_link() ->
+    pg:start_link(?MODULE).
+
+% Join a replication job pid to a RepId group
+%
+join({_, _} = RepId, Pid) when is_pid(Pid) ->
+    pg:join(?MODULE, id(RepId), Pid).
+
+% Leave a replication RepId group. This doesn't have to be called explicitly as
+% the processes are monitored and automatically removed by pg. It may be nice,
+% to call it from terminate/2 to speed things along a bit and clear the group
+% quicker.
+%
+leave({_, _} = RepId, Pid) when is_pid(Pid) ->
+    pg:leave(?MODULE, id(RepId), Pid).
+
+% Determine if a replication job should start on a particular node. If it
+% should, return `yes`, otherwise return `{no, OtherPid}`. `OtherPid` is
+% the pid of the replication job that is already running.
+%
+should_start({_, _} = RepId, Node) when is_atom(Node) ->
+    no_other_nodes(Node, pids(RepId)).
+
+% Determine if the replication job should keep running as the main job for that
+% RepId. If it is, return yes, otherwise return `{no, OtherPid}`. `OtherPid` is
+% the pid of the replication job that should stay running instead of this one.
+%
+should_run({_, _} = RepId, Pid) when is_pid(Pid) ->
+    case pids(RepId) of
+        [OtherPid | _] when OtherPid =/= Pid -> {no, OtherPid};
+        _ -> yes

Review Comment:
   Is there a reason to use `yes` here rather than the usual`true`?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: notifications-unsubscribe@couchdb.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org