)]}'
{"ansible/roles/rabbitmq/templates/rabbitmq.conf.j2":[{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"43936eb7f8d0daaddfe3d8619222cdb983f40988","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"889356d4_cc83e4fc","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"updated":"2021-05-21 05:04:09.000000000","message":"We\u0027ve been there and had this discussion in the past - and I don\u0027t think we should be doing that. Production architecture guide states deploying 2 rabbitmq nodes does not provide by default any HA and recommended is at least 3.","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"66b04a58c98bf090cc8dbbe38cf9e843cd97fb2b","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"daf03d02_a30821b9","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"281a17bc_986ee78b","updated":"2021-05-21 08:18:54.000000000","message":"split brain bad","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":27339,"name":"Michal Arbet","email":"michal.arbet@ultimum.io","username":"michalarbet"},"change_message_id":"3ee1365b9eefcba05ab8be06f42467bda8d784c1","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"884440ba_efbc2ded","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"5b64dac8_b994bf48","updated":"2021-05-21 08:29:20.000000000","message":"Done","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"67857f9dca55875f032e409bacb0a1666307b402","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"281a17bc_986ee78b","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"5d4b7787_32b5b743","updated":"2021-05-21 05:54:49.000000000","message":"Sure, but autoheal can hide real issues - and users can set it by themselves. I feel it should stay that way - we might add that to docs, but I don\u0027t think it should be \u0027\u0027out of the box\u0027\u0027.","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":27339,"name":"Michal Arbet","email":"michal.arbet@ultimum.io","username":"michalarbet"},"change_message_id":"6494a1969b0c50207188db1397eae10a7f9c8ed2","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"5d4b7787_32b5b743","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"889356d4_cc83e4fc","updated":"2021-05-21 05:47:29.000000000","message":"I know that per documentation is recommended to have three rabbitmq servers, but it should not be requirement to run kolla environment. I\u0027m sure that deployments with two rabbitmqs exists and  pause_minority stops whole cluster when one rabbit is down  while autoheal not.\n\nSo, before this option was changed, it was OK to run two rabbitmqs and when one was down, everything was OK, now it\u0027s not possible. Neither you nor I know how many such environments we have, but anyway, if conditional for such a case is not terrible and it makes sense I think.\n\nI agree with you that 3 nodes is better way, I also agree that pause_minority is better option to avoid split-brain. Just saying that there should be conditional for 2-nodes deployment.","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":27339,"name":"Michal Arbet","email":"michal.arbet@ultimum.io","username":"michalarbet"},"change_message_id":"3f551ddc44c90753aa6937e2903f7bd70888d5fa","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"e0ac84df_0d10b624","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"daf03d02_a30821b9","updated":"2021-05-21 08:22:29.000000000","message":"Could we make this configurable ? I mean, place rabbitmq_cluster_partition_handling variable to rabbitmq/defaults.yml and refer to it in rabbitmq\u0027s config template ?","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":30491,"name":"Radosław Piliszek","display_name":"Radek","email":"radek@piliszek.it","username":"yoctozepto","status":"self-employed techologist, collaborating mostly with 7bulls.com"},"change_message_id":"4b7322ead36a3ddb6fe8eae12cd0bc4fe323ed42","unresolved":true,"context_lines":[{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{% if groups.rabbitmq|length \u003d\u003d 2 %}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d autoheal"},{"line_number":11,"context_line":"{% else %}"},{"line_number":12,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":13,"context_line":"cluster_partition_handling \u003d pause_minority"}],"source_content_type":"text/x-jinja2","patch_set":1,"id":"5b64dac8_b994bf48","line":10,"range":{"start_line":10,"start_character":0,"end_line":10,"end_character":37},"in_reply_to":"e0ac84df_0d10b624","updated":"2021-05-21 08:25:50.000000000","message":"Ok.","commit_id":"72bb1e267ce10d23a4507fe54734177b61f76261"},{"author":{"_account_id":30491,"name":"Radosław Piliszek","display_name":"Radek","email":"radek@piliszek.it","username":"yoctozepto","status":"self-employed techologist, collaborating mostly with 7bulls.com"},"change_message_id":"088d3ea42ad43f865796655abe31bf18dd447d73","unresolved":true,"context_lines":[{"line_number":6,"context_line":"{% else %}"},{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d {{ rabbitmq_cluster_partition_handling }}"},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"management.listener.ip \u003d {{ api_interface_address }}"}],"source_content_type":"text/x-jinja2","patch_set":3,"id":"0ccbefda_0c91f2ce","line":9,"range":{"start_line":9,"start_character":3,"end_line":9,"end_character":29},"updated":"2021-05-21 08:33:43.000000000","message":"move this to the defaults main.yml as it does not make sense here","commit_id":"16f9056ba4a084fa08d454de24570b966960a793"},{"author":{"_account_id":27339,"name":"Michal Arbet","email":"michal.arbet@ultimum.io","username":"michalarbet"},"change_message_id":"b4a07d28a80f9f9f754633c8cd070fc808546c10","unresolved":false,"context_lines":[{"line_number":6,"context_line":"{% else %}"},{"line_number":7,"context_line":"listeners.tcp.1 \u003d {{ api_interface_address }}:{{ role_rabbitmq_port }}"},{"line_number":8,"context_line":"{% endif %}"},{"line_number":9,"context_line":"{# NOTE: to avoid split-brain #}"},{"line_number":10,"context_line":"cluster_partition_handling \u003d {{ rabbitmq_cluster_partition_handling }}"},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"management.listener.ip \u003d {{ api_interface_address }}"}],"source_content_type":"text/x-jinja2","patch_set":3,"id":"0f6a0bec_cee03cf5","line":9,"range":{"start_line":9,"start_character":3,"end_line":9,"end_character":29},"in_reply_to":"0ccbefda_0c91f2ce","updated":"2021-05-21 08:55:52.000000000","message":"Done","commit_id":"16f9056ba4a084fa08d454de24570b966960a793"}]}