)]}'
{"/COMMIT_MSG":[{"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":"3669a661a7d40e4000efd6445bc0c4cd5c8576e2","unresolved":true,"context_lines":[{"line_number":9,"context_line":"This patch starts using RBD active-active which is Cinder team recommended approach for grouping volume hosts that share the same backend conigurations to work in HA Active-Active mode."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Change-Id: I02817963c691784c126c79e330981e0cc771bc88"},{"line_number":12,"context_line":"Closes-Bug: #1904062"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"85fe8ba0_6f63ea22","line":12,"range":{"start_line":12,"start_character":0,"end_line":12,"end_character":6},"updated":"2022-06-23 15:07:09.000000000","message":"Related\n\nIt\u0027s not really fixing it. It\u0027s too late to fix it there. We just need to report back to ops how to fix it.","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"c89caa436f5917e1825ba7c86b1646e75b1702da","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This patch starts using RBD active-active which is Cinder team recommended approach for grouping volume hosts that share the same backend conigurations to work in HA Active-Active mode."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Change-Id: I02817963c691784c126c79e330981e0cc771bc88"},{"line_number":12,"context_line":"Closes-Bug: #1904062"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"cbde2c3f_1e856bad","line":12,"range":{"start_line":12,"start_character":0,"end_line":12,"end_character":6},"in_reply_to":"85fe8ba0_6f63ea22","updated":"2024-02-19 15:08:13.000000000","message":"Acknowledged","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"}],"/PATCHSET_LEVEL":[{"author":{"_account_id":30523,"name":"Dincer Celik","email":"hello@dincercelik.com","username":"osmanlicilegi"},"change_message_id":"bde4df7e41cfd8e186cd7b6a72e277d72198266b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"12f282de_1c059572","updated":"2021-11-22 08:19:21.000000000","message":"What is the situation of this change? I am using this at production for a long while with no issues. Any others?","commit_id":"091d40357d3363d46d6212ac3de9dfa042d09e4f"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"20282a070a1ada6d6438588e2ff9639390ccc2d4","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"980b6062_2076690d","updated":"2022-06-19 12:00:44.000000000","message":"From https://lists.openstack.org/pipermail/openstack-discuss/2020-November/018853.html the coordination backend isn\u0027t required, even just adding the cluster configuration will help. You could set the skip_backend_check variable and be happy if you want.\n\nThe two things that are still unclear to me:\n\n1. Why would we only want to this for Ceph? afaict all shared Cinder backends are affected.\n2. Can we handle upgrades better? i.e. in some automated fashion?","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"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":"e11dfcf8037e5daecc550a2f3dd91918f2c3b24b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"107a2c18_4f517988","updated":"2022-06-20 19:02:47.000000000","message":"Let\u0027s also finish testing the interaction with https://review.opendev.org/c/openstack/kolla-ansible/+/767951 first before approving.","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"author":{"_account_id":32398,"name":"Gaël THEROND","display_name":"Fl1nt","email":"gael.therond@bitswalk.com","username":"Fl1nt"},"change_message_id":"22600b919de8ed0be031c0cdd40495b2ef5b2f7d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"dd0aa2b1_b6ca7e48","updated":"2022-06-19 10:50:02.000000000","message":"We should definitely allow for fallback support.\nSome companies aren’t willing to use additional solutions such as the required additional K/V store.\n\nAdditionally, as already discussed on IRC, kolla purpose isn’t to enforce opinions on platforms.\n\nAs such, as cinder active/active isn’t enforced by Openstack itself, it should be an opt-in option as for the rolling upgrades already are.\n\nOh and we could also provide a way for ops to opt-in using memcached as K/V store for coordination as Tooz allow its usage.","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"author":{"_account_id":14200,"name":"Maksim Malchuk","email":"maksim.malchuk@gmail.com","username":"mmalchuk"},"change_message_id":"5f8c56e3b925a1ad981baf143a6827fedf66ba89","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"aa2e32a2_2d93a287","updated":"2023-12-15 10:49:17.000000000","message":"Sven, this change doesn\u0027t actually start using active-active for rbd, this achieved mostly by the inventory configuration and using the same name of the backend on each cluster node. and whatever redis/etcd coordination neither help with issue 1904062 nor help with active-active configuration already used.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":16198,"name":"Ilya Popov","email":"hebulrih@gmail.com","username":"IPO"},"change_message_id":"f18596ed85e94a0f2b726aceef0b048a34270257","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"4b52587d_e6ed877d","updated":"2023-06-02 06:57:45.000000000","message":"There is one issue we faced when performed tests fo cinder Active-Active with cluster conf:\neach cinder volume after start calculate and keep in memory allocated_capacity_gb. And report this value to cinder scheduler. Each time instance of cinder volume get task to create or delete volume - it increase or decrease allocated_capacity_gb accordingly. But this value isn\u0027t the same on another instances of cinder volume. And this different (and incorrect) values are reported to cinder scheduler. It leads to incorrect behaviour based in capacity - like decision about free space on backend to create additional volume.\n\nThe same situation was seen in \"pseudo-cluster config\" with backend_host parameter set.\n\nhttps://bugs.launchpad.net/cinder/+bug/1927186","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":14200,"name":"Maksim Malchuk","email":"maksim.malchuk@gmail.com","username":"mmalchuk"},"change_message_id":"1ddc2b9258314f8cf987b52051d5788f59b98c10","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"783ca251_0be53f50","updated":"2023-08-26 20:32:31.000000000","message":"agree. we shouldn\u0027t enforce. also this would break some current deploys.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"b39861f2febadfde1deaac0b600ce474f3d6fcc0","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"0e32f940_f011b081","updated":"2023-12-15 10:14:04.000000000","message":"please rebase and resolve merge conflict in the precheck.\n\nLGTM overall, I think most of the criticsm is moot or needs at the very least a more clear explanation what the problem actually is.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"eeba7e01c1683925754668a49b33f6eeb82ddd7c","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"898541d7_4f252c2a","in_reply_to":"25c1604e_98e99cc3","updated":"2025-03-20 16:34:40.000000000","message":"Done","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":23084,"name":"Bartosz Bezak","email":"bartosz@stackhpc.com","username":"b.bezak"},"change_message_id":"f4eec0a77c5e4a0003a4dac1b33bf6898a25ab8e","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"8810e6f9_8c52c5e4","in_reply_to":"2e38ad87_4b34fbba","updated":"2023-07-13 09:35:31.000000000","message":"I\u0027m wondering about real impact of this miscalculation of allocated_capacity_gb. By the look of it, the scheduler will incorrectly schedule volumes to less busy cinder-volume service. However the free_capacity_gb will still be reported correctly. So there is no risk of overcommit the backend. Therefore for setup with one backend managed by multiple HA cinder-volume service that is not a huge issue?\n\nFurthermore quota management is also probably not impacted here as well.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"9394c261e341205cd5cf7f425b15a8e89a6662bc","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"2e38ad87_4b34fbba","in_reply_to":"4b52587d_e6ed877d","updated":"2023-06-20 06:58:21.000000000","message":"That\u0027s probably still better than what we impose by default today ;-)","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"226bb9bfa2a3862e755e9db56b0aba43c3d5e706","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"fc79b524_87458c57","in_reply_to":"75ab5ad9_bcd3a924","updated":"2025-04-09 02:55:59.000000000","message":"Acknowledged","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"b643b6dc5bd0c6fc86c866c4c87c36d178cc82e8","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"f7b00432_e08dc4a4","in_reply_to":"8810e6f9_8c52c5e4","updated":"2023-11-02 10:49:51.000000000","message":"I actually don\u0027t think that a bug/edgecase in cinder itself should stop us from implementing this, we could point it out in the docs and upgrade notes, if we want to be extra nice.\n\nafaik other deployment tools already enable active-active HA (at least tripleO did imho) and cinder does also not say \"do not use active-active, because it is broken\".\n\nso we should not be perfect be the enemy of good here, imho and move on.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":14200,"name":"Maksim Malchuk","email":"maksim.malchuk@gmail.com","username":"mmalchuk"},"change_message_id":"024b80002a616ebe3964d7dbef176526324a32f2","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"072344e5_aec4caa3","in_reply_to":"898541d7_4f252c2a","updated":"2025-04-02 17:30:45.000000000","message":"thanks","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"4835dedf7468295b03cf4e6ca5d180543e45e189","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"25c1604e_98e99cc3","in_reply_to":"aa2e32a2_2d93a287","updated":"2024-02-19 16:11:56.000000000","message":"then the commit message is wrong, because it states:\n\n\"cinder: start using active-active for rbd\n\nThis patch starts using RBD active-active[...]\"\n\ncan we adjust the commit messages and the code, please? They really should tell what is done why. This commit message actually confused me.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"224c46f437e915bb2e529a20b60168325b02bb26","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"12efa4b1_bf27fcb0","in_reply_to":"aa2e32a2_2d93a287","updated":"2024-02-19 16:08:06.000000000","message":"then the commit message is wrong, because it states:\n\ncinder: start using active-active for rbd\n\nThis patch starts using RBD active-active\n\ncan we adjust the commit messages and the code, please? They really should tell what is done why. This commit message actually confused me.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":14200,"name":"Maksim Malchuk","email":"maksim.malchuk@gmail.com","username":"mmalchuk"},"change_message_id":"ea1c322540bfd31d8e6e3b99013c5d48f313be3a","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":7,"id":"75ab5ad9_bcd3a924","in_reply_to":"f7b00432_e08dc4a4","updated":"2023-11-02 11:27:11.000000000","message":"agree, but this is close-bug change as I can see, we can assume that this is something like workaround, mention about miscalculations in the docs, and merge it as related-bug change.","commit_id":"da3e9f7d0591d84a0a90e0a395f332d8a1e63fef"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"982e6ec8dfe2f8671722a22a1078c78bf4561c12","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":9,"id":"29421108_69b945cb","updated":"2024-02-19 16:13:35.000000000","message":"sorry for the comment sent twice, gerrit spit error messages at me and somehow this ended posting the comment twice.","commit_id":"59f3e9b70b8be2f438dfcc21d15fb70f7d424771"},{"author":{"_account_id":23084,"name":"Bartosz Bezak","email":"bartosz@stackhpc.com","username":"b.bezak"},"change_message_id":"a304a40083c61b0f3a1c7692d5fd16df8add2cec","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":15,"id":"4bc6f03c_ed50bcaa","updated":"2025-04-03 06:22:25.000000000","message":"recheck - error on pull","commit_id":"189943905a123ef259f4aa68b90ffe004bea1ec3"},{"author":{"_account_id":35263,"name":"Matt Crees","email":"mattc@stackhpc.com","username":"mattcrees"},"change_message_id":"d99ef872b5be58a15fcb69550ffd84c4a00711c5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":15,"id":"d918a6e7_3d8e15cb","updated":"2025-04-02 13:59:58.000000000","message":"recheck: `Internal Server Error (\\\"Get \\\"https://mirror-int.dfw.rax.opendev.org:4447/v2/\\\": net/http: TLS handshake timeout\\\")`","commit_id":"189943905a123ef259f4aa68b90ffe004bea1ec3"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"be3aea04446cbabefeccbb7020869ee4a1700bd2","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":18,"id":"cf98e420_5f08ca4b","updated":"2025-05-09 10:19:47.000000000","message":"let\u0027s get this in for epoxy","commit_id":"193830ef0f478a69b06365c9060b3da08f11c127"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"334c0f51ddb2a4bad555dd6965f534669c56e574","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":18,"id":"b8859608_2e874be2","updated":"2025-05-08 18:27:42.000000000","message":"recheck new images","commit_id":"193830ef0f478a69b06365c9060b3da08f11c127"}],"ansible/roles/cinder/tasks/precheck.yml":[{"author":{"_account_id":14200,"name":"Maksim Malchuk","email":"maksim.malchuk@gmail.com","username":"mmalchuk"},"change_message_id":"36f0154040885e99c1d08570bfce4ddd51f300da","unresolved":true,"context_lines":[{"line_number":63,"context_line":"  when:"},{"line_number":64,"context_line":"    - not skip_cinder_backend_check | bool"},{"line_number":65,"context_line":"    - cinder_backend_ceph | bool"},{"line_number":66,"context_line":"    - cinder_coordination_backend \u003d\u003d \u0027\u0027"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"- name: Check if S3 configurations are defined"},{"line_number":69,"context_line":"  assert:"}],"source_content_type":"text/x-yaml","patch_set":13,"id":"b66f72c6_99a44e4a","line":66,"updated":"2025-01-10 08:09:16.000000000","message":"should we check HA enabled here too. eg number of controllers \u003e1","commit_id":"bb393e4b694d27a27a35df67ae6f5566acc73aa4"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"eeba7e01c1683925754668a49b33f6eeb82ddd7c","unresolved":false,"context_lines":[{"line_number":63,"context_line":"  when:"},{"line_number":64,"context_line":"    - not skip_cinder_backend_check | bool"},{"line_number":65,"context_line":"    - cinder_backend_ceph | bool"},{"line_number":66,"context_line":"    - cinder_coordination_backend \u003d\u003d \u0027\u0027"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"- name: Check if S3 configurations are defined"},{"line_number":69,"context_line":"  assert:"}],"source_content_type":"text/x-yaml","patch_set":13,"id":"5a1965d2_ed5451b9","line":66,"in_reply_to":"b66f72c6_99a44e4a","updated":"2025-03-20 16:34:40.000000000","message":"Done","commit_id":"bb393e4b694d27a27a35df67ae6f5566acc73aa4"}],"ansible/roles/cinder/templates/cinder.conf.j2":[{"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":"3669a661a7d40e4000efd6445bc0c4cd5c8576e2","unresolved":true,"context_lines":[{"line_number":21,"context_line":"glance_ca_certificates_file \u003d {{ openstack_cacert }}"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"{% if cinder_backend_ceph | bool %}"},{"line_number":24,"context_line":"cluster \u003d ceph"},{"line_number":25,"context_line":"{% endif %}"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"{% if cinder_enabled_backends %}"}],"source_content_type":"text/x-jinja2","patch_set":6,"id":"f44a99d5_190726c2","line":24,"updated":"2022-06-23 15:07:09.000000000","message":"2 things to handle:\n\n1) It should only be set for ceph-volume.\n2) It should be set per cluster. Let\u0027s not assume that all cinder-volumes belong to the same cluster and instead let users decide per host. Do not get me wrong: I think that most of the time it will be the same cluster (with Ceph only at least). However, let\u0027s not wreak havoc if it is not the case. Also, notice this turns on clustering for *all backends*. I.e., enabling Ceph suddenly makes other backends try active-active HA only to fail (see https://review.opendev.org/c/openstack/kolla-ansible/+/847352). Also, other backends than Ceph might benefit from this setting (netapp, pure). I suggest we only document the way to get active-active HA and not force it on users at all. WDYT?","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"author":{"_account_id":22629,"name":"Michal Nasiadka","email":"mnasiadka@gmail.com","username":"mnasiadka"},"change_message_id":"eeba7e01c1683925754668a49b33f6eeb82ddd7c","unresolved":false,"context_lines":[{"line_number":21,"context_line":"glance_ca_certificates_file \u003d {{ openstack_cacert }}"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"{% if cinder_backend_ceph | bool %}"},{"line_number":24,"context_line":"cluster \u003d ceph"},{"line_number":25,"context_line":"{% endif %}"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"{% if cinder_enabled_backends %}"}],"source_content_type":"text/x-jinja2","patch_set":6,"id":"3c4170a1_45cf26de","line":24,"in_reply_to":"ab50b730_56ae3bdd","updated":"2025-03-20 16:34:40.000000000","message":"Done","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"},{"author":{"_account_id":32553,"name":"Sven Kieske","email":"sven_oss@posteo.de","username":"skieske"},"change_message_id":"b39861f2febadfde1deaac0b600ce474f3d6fcc0","unresolved":true,"context_lines":[{"line_number":21,"context_line":"glance_ca_certificates_file \u003d {{ openstack_cacert }}"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"{% if cinder_backend_ceph | bool %}"},{"line_number":24,"context_line":"cluster \u003d ceph"},{"line_number":25,"context_line":"{% endif %}"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"{% if cinder_enabled_backends %}"}],"source_content_type":"text/x-jinja2","patch_set":6,"id":"ab50b730_56ae3bdd","line":24,"in_reply_to":"f44a99d5_190726c2","updated":"2023-12-15 10:14:04.000000000","message":"\u003e  I suggest we only document the way to get active-active HA and not force it on users at all. WDYT?\n\nno, imho cinder should be HA by default, improving operational reliability when once cinder server fails.\n\nwe use the canonical cluster name \"ceph\" as well in other places, see:\n\nhttps://github.com/openstack/kolla-ansible/blob/583f884f1a6fe216b2fa84dd64beb63f1616c647/ansible/roles/cinder/defaults/main.yml#L270-L271\n\nI don\u0027t really understand 1), can you expand on that?\n\nregarding 2):\n\n\u003e  Let\u0027s not assume that all cinder-volumes belong to the same cluster and instead let users decide per host\n\ndo you meant to say let users decide per cluster?\n\nI also disagree to make this change larger by adding other storage backends.\n\nplease provide separate patches for other backends if your care about them.\n\nthis keeps git history clean and we can easily revert things if problems arise only in certain backend configurations.\n\nfor the record I\u0027m very much in favor to add also support for other backends, just not in the same changeset.","commit_id":"d72500c6ff6c378e044c381edc9c8a0d32159856"}]}