)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":23717,"name":"Surya Prakash Singh","email":"singh.surya64mnnit@gmail.com","username":"SuryaPrakash"},"change_message_id":"358e4084a0d9ea721f2ef6e73de7ed2dccb95b33","unresolved":false,"context_lines":[{"line_number":10,"context_line":"MariaDB backups might be handled."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Change-Id: I65f1ab92b9dce48cdd752fffd2123bd58d65b98f"},{"line_number":13,"context_line":"bp: https://blueprints.launchpad.net/kolla/+spec/database-backup-recovery"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"3f79a3b5_9ca0e244","line":13,"range":{"start_line":13,"start_character":0,"end_line":13,"end_character":73},"updated":"2018-08-31 06:11:45.000000000","message":"can you please follow this convention in writing message \nlike : bp: -\u003e  Implements:  \nhttps://wiki.openstack.org/wiki/GitCommitMessages","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"}],"specs/mariadb-backup-recovery.rst":[{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"a3c48042a977d5750e81e4926e67da92ff1fc25b","unresolved":false,"context_lines":[{"line_number":24,"context_line":"operator to be able to easily perform a restore and test the validity of their"},{"line_number":25,"context_line":"data."},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"As stated in the BP, general backup strategy should be considered out-of-scope"},{"line_number":28,"context_line":"as it\u0027s likely that each individual or organisation deploying OpenStack will"},{"line_number":29,"context_line":"have their own opinion on what should be done and the frequency with which"},{"line_number":30,"context_line":"these things should be performed.  However, Kolla should at least offer a way"},{"line_number":31,"context_line":"to expose the necessary mechanisms to facilitate existing strategies."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Use cases"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_bfec0e29","line":30,"range":{"start_line":27,"start_character":21,"end_line":30,"end_character":33},"updated":"2018-06-14 01:17:16.000000000","message":"If we do it in cronjob way, the frequency and full/incremetal may need to be considered, otherwise cronjob will be meanless. Maybe let operator to configure the backup frequency and backup rotate frequency.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"a3c48042a977d5750e81e4926e67da92ff1fc25b","unresolved":false,"context_lines":[{"line_number":48,"context_line":"Proposed change"},{"line_number":49,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There are several considerations as part of this proposed change.  There\u0027s the"},{"line_number":52,"context_line":"tooling necessary to perform a backup, the ability to schedule backups, and the"},{"line_number":53,"context_line":"requirement to transfer the data elsewhere."},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"Backup Tooling"},{"line_number":56,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_7ff2f606","line":53,"range":{"start_line":51,"start_character":67,"end_line":53,"end_character":43},"updated":"2018-06-14 01:17:16.000000000","message":"One way is transfer it to elsewhere. or maybe we do it on all mariadb node(typically three controllers).","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"59ae751c68f7e70af9696375c7c6d017119337e6","unresolved":false,"context_lines":[{"line_number":63,"context_line":"can have a significant performance impact when taking backups in a way that"},{"line_number":64,"context_line":"doesn\u0027t lock the database for the duration."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Instead, this proposed change will make use of Percona\u0027s XtraBackup tool, which"},{"line_number":67,"context_line":"has been designed specifically for \u0027hot-backups\u0027 avoiding locking and heavy"},{"line_number":68,"context_line":"performance impact. Because of the way XtraBackup functions, it also"},{"line_number":69,"context_line":"facilitates a simpler test / restore procedure as these are physical copies of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_eb68daf4","line":66,"range":{"start_line":66,"start_character":47,"end_line":66,"end_character":54},"updated":"2018-06-11 08:08:00.000000000","message":"IIRC percona packages are in EPEL repositories only in RHEL based distros. We are trying to remove epel from repositories for only use RDOs and other centos official repos","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":16984,"name":"Nick Jones","email":"nick@dischord.org","username":"yankcrime"},"change_message_id":"0ea9b3358d45511527d72d67af9d3db8f5897ac6","unresolved":false,"context_lines":[{"line_number":63,"context_line":"can have a significant performance impact when taking backups in a way that"},{"line_number":64,"context_line":"doesn\u0027t lock the database for the duration."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Instead, this proposed change will make use of Percona\u0027s XtraBackup tool, which"},{"line_number":67,"context_line":"has been designed specifically for \u0027hot-backups\u0027 avoiding locking and heavy"},{"line_number":68,"context_line":"performance impact. Because of the way XtraBackup functions, it also"},{"line_number":69,"context_line":"facilitates a simpler test / restore procedure as these are physical copies of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3f79a3b5_78f92b4b","line":66,"range":{"start_line":66,"start_character":47,"end_line":66,"end_character":54},"in_reply_to":"5f7c97a3_eb68daf4","updated":"2018-08-09 16:53:48.000000000","message":"In that case we could build the images using packages from Percona\u0027s repositories.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"59ae751c68f7e70af9696375c7c6d017119337e6","unresolved":false,"context_lines":[{"line_number":74,"context_line":"hosting the XtraBackup binary plus dependencies necessary to be able to"},{"line_number":75,"context_line":"connect to MariaDB and retrieve data from some or all of the databases."},{"line_number":76,"context_line":""},{"line_number":77,"context_line":"A Kolla-Ansible role will be created which will define tasks to:"},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"* Establish the necessary backup-specific credentials;"},{"line_number":80,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_0b5d1678","line":77,"range":{"start_line":77,"start_character":2,"end_line":77,"end_character":20},"updated":"2018-06-11 08:08:00.000000000","message":"this config could be integrated into same mariadb role with a variable to enable it","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"a3c48042a977d5750e81e4926e67da92ff1fc25b","unresolved":false,"context_lines":[{"line_number":74,"context_line":"hosting the XtraBackup binary plus dependencies necessary to be able to"},{"line_number":75,"context_line":"connect to MariaDB and retrieve data from some or all of the databases."},{"line_number":76,"context_line":""},{"line_number":77,"context_line":"A Kolla-Ansible role will be created which will define tasks to:"},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"* Establish the necessary backup-specific credentials;"},{"line_number":80,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_1fdd026f","line":77,"range":{"start_line":77,"start_character":2,"end_line":77,"end_character":20},"in_reply_to":"5f7c97a3_0b5d1678","updated":"2018-06-14 01:17:16.000000000","message":"Agree, it could be in mariadb role.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"f7390859df410c23bae2cb7b7e0a05fef04ea03e","unresolved":false,"context_lines":[{"line_number":80,"context_line":""},{"line_number":81,"context_line":"* Start a container from this image with an associated volume and perform a"},{"line_number":82,"context_line":"  full backup if no previous data exists, or an incremental backup if there is"},{"line_number":83,"context_line":"  existing data."},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The backup data will reside in a dedicated Docker volume.  This can then be"},{"line_number":86,"context_line":"used to facilitate transfer of the data elsewhere (i.e mounted by another"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_22a03ef4","line":83,"updated":"2018-06-10 08:50:48.000000000","message":"There would be some risk if we only have one full backup data, and all others are incremental backup. if one of the incremetal backup has problem, all later backup will be useless.\n\nWe are using XtraBackup in our Environment as Cronjob on each of the controller nodes. Our backup plan is(Just a sample):\n1. we do one full backup every one day.\n2. do incremetal backup every one hour.\n3. Keep full backup for two weeks, keep incremetal backup for one day.\n\n\nDo we consider to use \"cronjob\" way to instead of ansible playbook? \n\nThanks!","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"59ae751c68f7e70af9696375c7c6d017119337e6","unresolved":false,"context_lines":[{"line_number":80,"context_line":""},{"line_number":81,"context_line":"* Start a container from this image with an associated volume and perform a"},{"line_number":82,"context_line":"  full backup if no previous data exists, or an incremental backup if there is"},{"line_number":83,"context_line":"  existing data."},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The backup data will reside in a dedicated Docker volume.  This can then be"},{"line_number":86,"context_line":"used to facilitate transfer of the data elsewhere (i.e mounted by another"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_cbd71ebf","line":83,"in_reply_to":"5f7c97a3_22a03ef4","updated":"2018-06-11 08:08:00.000000000","message":"Agree, a cron could be implemented in the container, check fernet-rotate.sh script in keystone role.similar behavior","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":16984,"name":"Nick Jones","email":"nick@dischord.org","username":"yankcrime"},"change_message_id":"0ea9b3358d45511527d72d67af9d3db8f5897ac6","unresolved":false,"context_lines":[{"line_number":80,"context_line":""},{"line_number":81,"context_line":"* Start a container from this image with an associated volume and perform a"},{"line_number":82,"context_line":"  full backup if no previous data exists, or an incremental backup if there is"},{"line_number":83,"context_line":"  existing data."},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The backup data will reside in a dedicated Docker volume.  This can then be"},{"line_number":86,"context_line":"used to facilitate transfer of the data elsewhere (i.e mounted by another"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3f79a3b5_db09592c","line":83,"in_reply_to":"5f7c97a3_22a03ef4","updated":"2018-08-09 16:53:48.000000000","message":"This seems like a sensible schedule to propose.  However, I\u0027m wary of shipping a default backup schedule as it raises a few questions...\n\nI would like to assume that most people would want to perform a backup (full or incremental) when the platform is under the least amount of load, i.e outside of \u0027normal\u0027 working hours.  So let\u0027s say that we decided to schedule them for midnight by default.  How do we know what the timezone is in order to ensure that 00:00 is truly midnight local time?  I\u0027m thinking about situations where the server\u0027s TZ might be UTC, but the business in question is UTC+8, for example.\n\nIt\u0027s for these reasons that I originally proposed that we just offer the mechanism, as opposed to also including the operational procedure.\n\nHappy to be convinced either way!","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":92,"context_line":"should be possible to nominate any node which has access to either the internal"},{"line_number":93,"context_line":"API address or the database node directly.  It\u0027s up to the operator to choose"},{"line_number":94,"context_line":"which mode is best for them, as there are a number of different considerations"},{"line_number":95,"context_line":"and trade-offs to make."},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"Scheduling"},{"line_number":98,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_2bc26d54","line":95,"updated":"2018-06-08 15:22:00.000000000","message":"Would this be a configuration option? With a default something like:\n\nmariadb_backup_host: \"{{ groups[\u0027mariadb\u0027][0] }}\"","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"a3c48042a977d5750e81e4926e67da92ff1fc25b","unresolved":false,"context_lines":[{"line_number":92,"context_line":"should be possible to nominate any node which has access to either the internal"},{"line_number":93,"context_line":"API address or the database node directly.  It\u0027s up to the operator to choose"},{"line_number":94,"context_line":"which mode is best for them, as there are a number of different considerations"},{"line_number":95,"context_line":"and trade-offs to make."},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"Scheduling"},{"line_number":98,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_bf3ace94","line":95,"in_reply_to":"5f7c97a3_2bc26d54","updated":"2018-06-14 01:17:16.000000000","message":"or maybe descovery one host which in primary state.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":16984,"name":"Nick Jones","email":"nick@dischord.org","username":"yankcrime"},"change_message_id":"0ea9b3358d45511527d72d67af9d3db8f5897ac6","unresolved":false,"context_lines":[{"line_number":92,"context_line":"should be possible to nominate any node which has access to either the internal"},{"line_number":93,"context_line":"API address or the database node directly.  It\u0027s up to the operator to choose"},{"line_number":94,"context_line":"which mode is best for them, as there are a number of different considerations"},{"line_number":95,"context_line":"and trade-offs to make."},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"Scheduling"},{"line_number":98,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3f79a3b5_bb0edd2f","line":95,"in_reply_to":"5f7c97a3_bf3ace94","updated":"2018-08-09 16:53:48.000000000","message":"Yes, or actually pick one which isn\u0027t the primary.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":98,"context_line":"----------"},{"line_number":99,"context_line":""},{"line_number":100,"context_line":"Automatic scheduling of backups should be disabled by default, but Kolla should"},{"line_number":101,"context_line":"provide a mechanism to facilitate this if it\u0027s an operational requirement."},{"line_number":102,"context_line":""},{"line_number":103,"context_line":"Archival"},{"line_number":104,"context_line":"--------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_0b9a116e","line":101,"updated":"2018-06-08 15:22:00.000000000","message":"Could mention the in-container vs. out-of-container tradeoff here. If kolla is to manage the scheduling, I think an in-container cron job (separate from the existing cron-tainer) is going to offer least friction.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":16984,"name":"Nick Jones","email":"nick@dischord.org","username":"yankcrime"},"change_message_id":"0ea9b3358d45511527d72d67af9d3db8f5897ac6","unresolved":false,"context_lines":[{"line_number":98,"context_line":"----------"},{"line_number":99,"context_line":""},{"line_number":100,"context_line":"Automatic scheduling of backups should be disabled by default, but Kolla should"},{"line_number":101,"context_line":"provide a mechanism to facilitate this if it\u0027s an operational requirement."},{"line_number":102,"context_line":""},{"line_number":103,"context_line":"Archival"},{"line_number":104,"context_line":"--------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3f79a3b5_7b184574","line":101,"in_reply_to":"5f7c97a3_0b9a116e","updated":"2018-08-09 16:53:48.000000000","message":"Yeah, but arguably at the expense of visibility.  I\u0027ll add some wording to that effect.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"59ae751c68f7e70af9696375c7c6d017119337e6","unresolved":false,"context_lines":[{"line_number":98,"context_line":"----------"},{"line_number":99,"context_line":""},{"line_number":100,"context_line":"Automatic scheduling of backups should be disabled by default, but Kolla should"},{"line_number":101,"context_line":"provide a mechanism to facilitate this if it\u0027s an operational requirement."},{"line_number":102,"context_line":""},{"line_number":103,"context_line":"Archival"},{"line_number":104,"context_line":"--------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_2bf01267","line":101,"in_reply_to":"5f7c97a3_0b9a116e","updated":"2018-06-11 08:08:00.000000000","message":"agree","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"c1dd28f11be8011e4a507414127bb7dbcfdd5936","unresolved":false,"context_lines":[{"line_number":104,"context_line":"--------"},{"line_number":105,"context_line":""},{"line_number":106,"context_line":"Another tool is required to manage the backup lifecycle.  This is currently"},{"line_number":107,"context_line":"considered out of scope."},{"line_number":108,"context_line":""},{"line_number":109,"context_line":"Backup Restore"},{"line_number":110,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_81b9e2c5","line":107,"updated":"2018-07-31 12:26:20.000000000","message":"just mount a docker volume and let operators make it redundant","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":16984,"name":"Nick Jones","email":"nick@dischord.org","username":"yankcrime"},"change_message_id":"0ea9b3358d45511527d72d67af9d3db8f5897ac6","unresolved":false,"context_lines":[{"line_number":104,"context_line":"--------"},{"line_number":105,"context_line":""},{"line_number":106,"context_line":"Another tool is required to manage the backup lifecycle.  This is currently"},{"line_number":107,"context_line":"considered out of scope."},{"line_number":108,"context_line":""},{"line_number":109,"context_line":"Backup Restore"},{"line_number":110,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3f79a3b5_1b291189","line":107,"in_reply_to":"5f7c97a3_81b9e2c5","updated":"2018-08-09 16:53:48.000000000","message":"+1","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":113,"context_line":"facilitating a restore is relatively straightforward.  Automating this is"},{"line_number":114,"context_line":"currently out of scope, but this piece of work should include an example"},{"line_number":115,"context_line":"procedure for how to handle this volume and access the data that\u0027s been backed"},{"line_number":116,"context_line":"up."},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"Security impact"},{"line_number":119,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_0b73f14f","line":116,"updated":"2018-06-08 15:22:00.000000000","message":"Would be good to see a workflow that includes rolling back to an archived backup, as well as the most recent.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":19316,"name":"Eduardo Gonzalez","email":"dabarren@gmail.com","username":"egonzalez90"},"change_message_id":"c1dd28f11be8011e4a507414127bb7dbcfdd5936","unresolved":false,"context_lines":[{"line_number":122,"context_line":"role within MariaDB in order to give the tooling the necessary access.  This"},{"line_number":123,"context_line":"will be read-only in nature and restricted to these specific privileges:"},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"``SELECT,RELOAD,LOCK TABLES,SHOW VIEW,REPLICATION CLIENT``"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"Performance Impact"},{"line_number":128,"context_line":"------------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_a1f65e8f","line":125,"range":{"start_line":125,"start_character":0,"end_line":125,"end_character":58},"updated":"2018-07-31 12:26:20.000000000","message":"Nice to see a security minded permissions :)","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":132,"context_line":"the backup target device is the same as the source."},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Aside from degradation incurred by way of I/O contention, the selection of"},{"line_number":135,"context_line":"XtraBackup is an attempt at mitigating any kind of performance impact."},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"Implementation"},{"line_number":138,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_ebe1d5d1","line":135,"updated":"2018-06-08 15:22:00.000000000","message":"Could we use some clever scheduling to back up each DB at a different time of day to ease the load? Or at least make this configurable, I guess some sites might want to backup everything overnight.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":9820,"name":"Liping Mao","email":"limao@cisco.com","username":"LipingMao"},"change_message_id":"a3c48042a977d5750e81e4926e67da92ff1fc25b","unresolved":false,"context_lines":[{"line_number":132,"context_line":"the backup target device is the same as the source."},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Aside from degradation incurred by way of I/O contention, the selection of"},{"line_number":135,"context_line":"XtraBackup is an attempt at mitigating any kind of performance impact."},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"Implementation"},{"line_number":138,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_1f22e252","line":135,"in_reply_to":"5f7c97a3_ebe1d5d1","updated":"2018-06-14 01:17:16.000000000","message":"Agree. Backup it in lower load time, and do a full backup per day, do incremetal backup base on the full backup of last day.\n\nWe tested the backup performance in 60 compute nodes env. our mysql is about 1G (we enable nova/neutron/keystone/glance basic service), and while backup we do rally test to call api. It takes about 1 min to do a full backup. and takes about 10+s to do incremetal backup.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":1297,"name":"Joshua Harlow","email":"jxharlow@godaddy.com","username":"harlowja"},"change_message_id":"40dad2cca27abfa4b06de6c8572229673024225e","unresolved":false,"context_lines":[{"line_number":147,"context_line":"Work Items"},{"line_number":148,"context_line":"----------"},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"1. Introduce a new Kolla image containing XtraBackup package plus dependencies"},{"line_number":151,"context_line":"   such as scripts to handle triggering the backup;"},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"2. Introduce a new Kolla-Ansible command and corresponding role to take a"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_4f30289f","line":150,"updated":"2018-06-07 20:13:51.000000000","message":"Seems like kolla-toolbox could house this?","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":147,"context_line":"Work Items"},{"line_number":148,"context_line":"----------"},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"1. Introduce a new Kolla image containing XtraBackup package plus dependencies"},{"line_number":151,"context_line":"   such as scripts to handle triggering the backup;"},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"2. Introduce a new Kolla-Ansible command and corresponding role to take a"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_0bef71b1","line":150,"in_reply_to":"5f7c97a3_4f30289f","updated":"2018-06-08 15:22:00.000000000","message":"While we could use the kolla-toolbox image, we couldn\u0027t use the kolla_toolbox container as the backup container needs access to the backup docker volume, and we wouldn\u0027t want to expose that inside kolla_toolbox. In that case, perhaps it makes sense to stick with a purpose-built image with less bloat.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":155,"context_line":"   dedicated volume;"},{"line_number":156,"context_line":""},{"line_number":157,"context_line":"3. Documentation for new options and also restore process, along with examples."},{"line_number":158,"context_line":""},{"line_number":159,"context_line":"Testing"},{"line_number":160,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":161,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_cb13b9b7","line":158,"updated":"2018-06-08 15:22:00.000000000","message":"An in-container scheduling approach (if we go down that path) would require some support in kolla and kolla ansible.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"},{"author":{"_account_id":14826,"name":"Mark Goddard","email":"markgoddard86@gmail.com","username":"mgoddard"},"change_message_id":"e807d5113a15c85a22a53d699cfe6435b1c445bd","unresolved":false,"context_lines":[{"line_number":162,"context_line":"Tests should be added to validate that a backup has been taken successfully"},{"line_number":163,"context_line":"with the default settings in place.  This would take the form of starting"},{"line_number":164,"context_line":"another MariaDB container with the backup volume mounted as ``/var/lib/mysql``"},{"line_number":165,"context_line":"and then performing some example queries to ensure expected data is returned."},{"line_number":166,"context_line":""},{"line_number":167,"context_line":"Documentation Impact"},{"line_number":168,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5f7c97a3_ab83ddca","line":165,"updated":"2018-06-08 15:22:00.000000000","message":"Kolla doesn\u0027t really have any tests like this currently, so I\u0027m not sure exactly where it would fit in. It would be great to see though, and we do need to start improving coverage of features like this.","commit_id":"d41553a7f0c0e154824460daf4a96659de20cd33"}]}