capistrant opened a new pull request, #12599: URL: https://github.com/apache/druid/pull/12599
<!-- Thanks for trying to help us make Apache Druid be the best it can be! Please fill out as much of the following information as is possible (where relevant, and remove it when irrelevant) to help make the intention and scope of this PR clear in order to ease review. --> <!-- Please read the doc for contribution (https://github.com/apache/druid/blob/master/CONTRIBUTING.md) before making this PR. Also, once you open a PR, please _avoid using force pushes and rebasing_ since these make it difficult for reviewers to see what you've changed in response to their reviews. See [the 'If your pull request shows conflicts with master' section](https://github.com/apache/druid/blob/master/CONTRIBUTING.md#if-your-pull-request-shows-conflicts-with-master) for more details. --> This PR addresses part of the proposal in #12526 <!-- Replace XXXX with the id of the issue fixed in this PR. Remove this section if there is no corresponding issue. Don't reference the issue in the title of this pull-request. --> <!-- If you are a committer, follow the PR action item checklist for committers: https://github.com/apache/druid/blob/master/dev/committer-instructions.md#pr-and-issue-action-item-checklist-for-committers. --> ### Description <!-- Describe the goal of this PR, what problem are you fixing. If there is a corresponding issue (referenced above), it's not necessary to repeat the description here, however, you may choose to keep one summary sentence. --> <!-- Describe your patch: what did you change in code? How did you fix the problem? --> <!-- If there are several relatively logically separate changes in this PR, create a mini-section for each of them. For example: --> #### Made an incompatible change to druid_segments table Added a new column, `last_used VARCHAR(255)` to the druid_segments table. When creating the druid_segments table from scratch, this column is not nullable. However, to more easily facilitate an upgrade to a version of Druid with this code, the migration steps to ALTER druid_segments allow the column to have null values. Perhaps we can document an optional post upgrade step to ALTER the table to not allow nulls that is contingent upon completing the existing optional post-upgrade step to populate last_used for already unused rows. This column is a UTC date string corresponding to the last time that the `used` column was modified. #### Added a configuration to the `druid.coordinator.kill.` family, `druid.coordinator.kill.bufferPeriod` `druid.coordinator.kill.bufferPeriod` is a Duration that defines the amount of time that a segment must have been unused for before KillUnusedSegments will potentially kill it. For example, with the default `PT24H`. If I mark a segment X unused at `2022-06-01T00:05:00.000Z`. By rule, this segment cannot be killed until at or after `2022-06-02T00:05:00.000Z`. The most prominent use of this new configuration is to prevent deletion of data from Druid by mistake. It can be thought of like the trash folder in HDFS. Marking segments unused is the `rm`. Without trash, the data is just gone. With trash, we can recover. The period is the amount of time before pending trash sent away for good. <!-- In each section, please describe design decisions made, including: - Choice of algorithms - Behavioral aspects. What configuration values are acceptable? How are corner cases and error conditions handled, such as when there are insufficient resources? - Class organization and design (how the logic is split between classes, inheritance, composition, design patterns) - Method organization and design (how the logic is split between methods, parameters and return types) - Naming (class, method, API, configuration, HTTP endpoint, names of emitted metrics) --> <!-- It's good to describe an alternative design (or mention an alternative name) for every design (or naming) decision point and compare the alternatives with the designs that you've implemented (or the names you've chosen) to highlight the advantages of the chosen designs and names. --> #### Alternative Design Alternatively, I could have embedded last_used within the payload. The functionality at the end of the day would be the same, but I have listed some pros and cons below. Pros * No table schema change needed to upgrade Druid, making life easier for an operator. Cons * If you were adding this feature in the initial pre-release development of Druid, you'd probably steer away from accumulating too much in the payload column and instead create native columns in the database table * embedding it in this column is almost hiding it. Or at least making it harder for others to discover and understand. * You need to pull the payload back in the query for unused segments. Increased db I/O, network, druid memory and compute demands * You need to process the payloads in Druid code instead of simply filtering on the column value if it were native to the metastore table #### Upgrade Plan Requiring a database table alteration is a breaking change. It introduces an extra requirement for an operator to upgrade from an older version of Druid. I have done my best to limit the complexity of upgrading and provide as much assistance as possible to the Druid operators out there. ##### No work necessary case There is a case where there is no extra work necessary for the Druid operator to upgrade. If the operator has used the default of `true` for `druid.metadata.storage.connector.createTables`, and their metastore user has DDL privs, the table will automatically be altered at startup if the `last_used` column is missing from the druid_segments table. The coordinator and overlord will startup just fine. And all future changes to the used column (plus new segment creation) will populate `last_used`. For existing segments who match `used = true`, the value of `last_used` will be `null`. I have coded up the logic for finding segments to kill in a way that ignores the bufferPeriod for these segments with `null` column values. This means that `KillUnusedSegments` will use the same logic as pre-upgrade when looking for killable segments in the metadata store. ##### Some work necessary case If the operator has either: * set `druid.metadata.storage.connector.createTables` to false * not given their metastore user DDL privs They will need to execute an ALTER statement themselves before the upgrade. To make things easy for them, I have created a new Druid cli tool called UpdateTables that can perform this for them by executing the same alter table code path as the no work necessary case described above. This tool, as well as the actual ALTER command - in case the operator wants to do it themselves - is documented in a new upgrade-prep.md resource. ##### Optional post upgrade action If the operator is interested in populating all of their legacy unused segments with a last_used date string following the upgrade, I have included another action in the new UpdateTables cli tool that will do so. Again, I have documented the tool and provided the actual UPDATE statement that they could use manually. This step is completely optional and only needed if the operator desires that their legacy unused segments come under the purview of the new bufferPeriod config post-upgrade. <!-- If there was a discussion of the design of the feature implemented in this PR elsewhere (e. g. a "Proposal" issue, any other issue, or a thread in the development mailing list), link to that discussion from this PR description and explain what have changed in your final design compared to your original proposal or the consensus version in the end of the discussion. If something hasn't changed since the original discussion, you can omit a detailed discussion of those aspects of the design here, perhaps apart from brief mentioning for the sake of readability of this PR description. --> <!-- Some of the aspects mentioned above may be omitted for simple and small changes. --> <hr> ##### Key changed/added classes in this PR * MetadataStorageConnector - added 2 methods to interface * IndexerSQLMetadataStorageCoordinator - create new segments with last_used column populated * SQLMetadataConnector - code for altering and validating the druid_segments table * SQLMetadataSegmentPublisher - publish segments with last_used column populated * SegmentsMetadataManager - update signature of interface method * SqlSegmentsMetadataManager - manage and use last_used column properly * SqlSegmentsMetadataQuery.java - manage last_used properly * DruidCoordinatorConfig - new config * Main - new cli tool * UpdateTables - new cli tool <hr> <!-- Check the items by putting "x" in the brackets for the done things. Not all of these items apply to every PR. Remove the items which are not done or not relevant to the PR. None of the items from the checklist below are strictly necessary, but it would be very helpful if you at least self-review the PR. --> This PR has: - [ ] been self-reviewed. - [ ] added documentation for new or modified features or behaviors. - [ ] added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links. - [ ] added or updated version, license, or notice information in [licenses.yaml](https://github.com/apache/druid/blob/master/dev/license.md) - [ ] added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader. - [ ] added unit tests or modified existing tests to cover new code paths, ensuring the threshold for [code coverage](https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md) is met. - [ ] added integration tests. - [ ] been tested in a test Druid cluster. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
