--- title: "Easysearch 跨版本兼容性测试,还原 Elasticsearch 各版本快照数据" date: 2023-06-05 lastmod: 2023-06-05 description: "本文测试了不同版本的 Elasticsearch 快照在 Easysearch 中的数据恢复能力,包括索引、别名、模版和生命周期策略的恢复,并通过数据比对验证一致性,结果显示多种版本均能成功恢复索引文档和别名。" tags: ["Easysearch", "Elasticsearch", "snapshot", "restore"] summary: "本文主要测试验证 Elasticsearch 各版本快照在 Easysearch 中进行数据恢复。 准备测试数据 # 索引 # 别名 # 模版 # 生命周期策略 # 创建快照 # PUT /_snapshot/my_backup { "type": "fs", "settings": { "location": "/infini/test/es_backup" } } PUT /_snapshot/my_backup/snapshot_1 { "indices": "*", "ignore_unavailable": false, "include_global_state": false } GET /_snapshot/my_backup/snapshot_1 ignore_unavailable:如果 indices 列表中的索引不存在,则是否忽略该索引而不是使快照失败。默认值为 false 。 include_global_state:是否在快照中包含集群状态(包括索引模版、生命周期配置、持久化配置等)。默认值为 true ,建议设为 false。 恢复快照 # POST /_snapshot/my_backup/snapshot_1/_restore { "indices": "*", "ignore_unavailable": false, "include_global_state": false, "include_aliases": true, "ignore_index_settings": [ "index." --- 本文主要测试验证 Elasticsearch 各版本快照在 Easysearch 中进行数据恢复。 ### 准备测试数据 #### 索引 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p1.png" "" %}} #### 别名 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p2.png" "" %}} #### 模版 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p3.png" "" %}} #### 生命周期策略 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p4.png" "" %}} ### 创建快照 ``` PUT /_snapshot/my_backup { "type": "fs", "settings": { "location": "/infini/test/es_backup" } } PUT /_snapshot/my_backup/snapshot_1 { "indices": "*", "ignore_unavailable": false, "include_global_state": false } GET /_snapshot/my_backup/snapshot_1 ``` - ignore_unavailable:如果 indices 列表中的索引不存在,则是否忽略该索引而不是使快照失败。默认值为 false 。 - include_global_state:是否在快照中包含集群状态(包括索引模版、生命周期配置、持久化配置等)。默认值为 true ,建议设为 false。 ### 恢复快照 ``` POST /_snapshot/my_backup/snapshot_1/_restore { "indices": "*", "ignore_unavailable": false, "include_global_state": false, "include_aliases": true, "ignore_index_settings": [ "index.lifecycle.indexing_complete" ] } ``` - ignore_unavailable:如果 indices 列表中的索引不存在,则是否忽略该索引而不是使还原操作失败。默认值为 false 。 - include_global_state:是否还原群集状态。默认值为 false 。 - include_aliases:是否恢复别名及其关联索引。默认值为 true 。 - index.lifecycle.indexing_complete 配置不支持,忽略掉。 ### 数据验证 #### 索引 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p5.png" "" %}} 通过 gateway 进行数据比对 ``` path.data: data path.logs: log #show progress bar #progress_bar.enabled: true elasticsearch: - name: source enabled: true endpoints: - http://192.168.3.185:29200 - name: target enabled: true endpoints: - https://192.168.3.185:9205 basic_auth: username: admin password: admin pipeline: - name: index_diff_service auto_start: true processor: - dag: mode: wait_all parallel: - dump_hash: #dump es1's doc sort_document_fields: true indices: ".infini_activities-000004" ##需要比对的索引名 scroll_time: "10m" elasticsearch: "source" # query_string: "_id:c8es70pu46lgfdgmja9g-1646117763293610802-2" # fields: "doc_hash" output_queue: "source_docs" batch_size: 5000 slice_size: 1 # hash_func: "xxhash64" - dump_hash: #dump es2's doc indices: ".infini_activities-000004" scroll_time: "10m" # fields: "doc_hash" # query_string: "_id:c8es70pu46lgfdgmja9g-1646117763293610802-2" batch_size: 5000 slice_size: 1 # hash_func: "xxhash64" elasticsearch: "target" output_queue: "target_docs" end: - index_diff: diff_queue: "diff_result" buffer_size: 10 text_report: true #如果要存 es,这个开关关闭,开启 pipeline 的 diff_result_ingest 任务 source_queue: "source_docs" target_queue: "target_docs" #pipeline: # - name: diff_result_ingest # processor: # - json_indexing: # index_name: "diff_result" # elasticsearch: "source" # input_queue: "diff_result" ``` ./gateway-linux-amd64 -config data_check.yml {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p6.png" "" %}} #### 别名 {{% load-img "/img/blog/2023/easysearch-snapshot-restore/p7.png" "" %}} #### 模版 ``` PUT _template/.infini_activities-rollover { "order": 100000, "index_patterns": [ ".infini_activities*" ], "settings": { "index": { "format": "7", "lifecycle": { "name": "ilm_.infini_metrics-30days-retention", "rollover_alias": ".infini_activities" }, "codec": "best_compression", "number_of_shards": "1", "translog": { "durability": "async" } } }, "mappings": { "dynamic_templates": [ { "strings": { "mapping": { "ignore_above": 256, "type": "keyword" }, "match_mapping_type": "string" } } ] }, "aliases": {} } PUT _template/.infini { "order": 0, "index_patterns": [ ".infini_*" ], "settings": { "index": { "max_result_window": "10000000", "mapping": { "total_fields": { "limit": "20000" } }, "analysis": { "analyzer": { "suggest_text_search": { "filter": [ "word_delimiter" ], "tokenizer": "classic" } } }, "number_of_shards": "1" } }, "mappings": { "dynamic_templates": [ { "strings": { "mapping": { "ignore_above": 256, "type": "keyword" }, "match_mapping_type": "string" } } ] }, "aliases": {} } ``` #### 生命周期策略 ``` PUT _ilm/policy/ilm_.infini_metrics-30days-retention { "policy": { "phases": { "hot": { "min_age": "0ms", "actions": { "rollover": { "max_size": "50gb", "max_age": "30d" }, "set_priority": { "priority": 100 } } }, "delete": { "min_age": "30d", "actions": { "delete": { } } } } } } ``` 注:不支持 "delete_searchable_snapshot": true 配置 ### 测试结果 | 源集群(Elasticsearch) | 目标集群(Easysearch) | 测试结果 | | ----------------------- | ---------------------- | -------------------------- | | 7.10.2 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.10.1 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.10.0 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.9.2 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.9.0 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.8.1 | 1.0.0 | 索引文档一致,别名恢复成功 | | 7.5.2 | 1.0.0 | 索引文档一致,别名恢复成功 | | 6.8.12 | 1.0.0 | 索引文档一致,别名恢复成功 | | 6.5.4 | 1.0.0 | 索引文档一致,别名恢复成功 |