In https://github.com/apache/couchdb/blob/master/src/rexi/src/rexi_sup.erl#L23 noticed it's one_for_one strategy, so if any of the top level children dies and restarts (like say rexi_sup did in the log), when rexi_sup respawns, nothing will restart all the rexi servers.
`rexi_server_mon` only checks if it needs to restart servers on cluster membership changes and on startup: https://github.com/apache/couchdb/blob/master/src/rexi/src/rexi_server_mon.erl#L56-L63 So maybe we'd want a `rest_for_one` strategy in rexi_sup (in addition to increasing the intensity check timeout in rexi_server_sup?). ``` init([]) -> {ok, {{rest_for_one, 3, 10}, [ { rexi_server, {rexi_server, start_link, [rexi_server]}, permanent, 100, worker, [rexi_server] }, { rexi_server_sup, {rexi_server_sup, start_link, [rexi_server_sup]}, permanent, 100, supervisor, [rexi_server_sup] }, { rexi_server_mon, {rexi_server_mon, start_link, [rexi_server]}, permanent, 100, worker, [rexi_server_mon] }, { rexi_buffer_sup, {rexi_server_sup, start_link, [rexi_buffer_sup]}, permanent, 100, supervisor, [rexi_server_sup] }, { rexi_buffer_mon, {rexi_server_mon, start_link, [rexi_buffer]}, permanent, 100, worker, [rexi_server_mon] } ]}}. ``` In this case if rexi_server_sup dies, it will down rexi_server_mon (well and unfortunately the buffers), but then on restart it will bring back rexi_server_mon and it should repopulate the servers in rexi_server_sup. [ Full content available at: https://github.com/apache/couchdb/issues/1571 ] This message was relayed via gitbox.apache.org for [email protected]
