generic().execute(new RecoveryRunner(recoveryId));任务处理模块:indices/recovery/RecoveryTarget.java 在doRecovery函数中,将本次要恢复的 shard 相关信息,如 shardid,metadataSnapshot 重要的是metadataSnapshot中包含 syncid等,封装成 StartRecoveryRequest ,RPC 发送出去:...
cluster.routing.allocation.node_concurrent_recoveries 单个节点最大并发进/出 recovery 数,默认2 indices.recovery.max_bytes_per_sec 默认40m indices.recovery.concurrent_streams 单个节点恢复时可以打开的网络流数量,默认3 cluster.routing.allocation.node_concurrent_recoveries/出recovery,默认2 indices.recovery.max...
indices.recovery.max_bytes_per_sec 默认40m indices.recovery.concurrent_streams 单个节点恢复时可以打开的网络流数量,默认3 即使关闭限速,这个阶段仍然可能非常漫长,目前最好的方式就是先执行 synced flush, 但是 syncd flush 并且本身也可能比较慢,因为我们常常为了优化写入速度而加大 translog 刷盘周期,也会延长...
In Elasticsearch,recoveryrefers to the process of recovering a shard when something goes wrong. Shard recoveries can take place in various circumstances, such as when a node fails and a replica shard needs to be recreated from a primary shard, when theclusterneeds to relocate shards to different...
importstaticorg.elasticsearch.indices.recovery.RecoverySettings.INDICES_RECOVERY_MAX_BYTES_PER_SEC_SETTING; importstaticorg.elasticsearch.test.NodeRoles.nonMasterNode; Expand DownExpand Up@@ -314,11 +315,15 @@ public void test3MasterNodes2Failed() throws Exception { ...