Switch (vDS) Health Check 354 Next Steps 355 Setting Up VMware vSphere DRS and HA 355 Enabling DRS 355 Enabling HA 356 Next Steps 359 Setting Up VMware vSphere Storage 359 Configuring vSAN 359 Upgrading the Service 386 Removing the Deprecated HA Service 386 Removing HA Service for a Cluster with a Single Segment per Segment Host 387 Removing HA Service for a Cluster with a Multiple Segments per Installing the Postmaster Service 388 Managing Single Host Failures in VMware vSphere 388 How DRS and HA protect a running Greenplum cluster 389 Single Host Failure Scenarios 389 Mirror Segment Failure

Switch (vDS) Health Check 337 Next Steps 338 Setting Up VMware vSphere DRS and HA 338 Enabling DRS 338 Enabling HA 339 Next Steps 342 Setting Up VMware vSphere Storage 342 Configuring vSAN 342 Cluster Connection Information 366 Managing Single Host Failures in VMware vSphere 366 How DRS and HA protect a running Greenplum cluster 367 Single Host Failure Scenarios 367 VMware Greenplum 6 Documentation Standby Master Failure 369 Master Failure 370 After Host Recovery 370 Best Practices for Greenplum HA 371 Appendix: Grace Period for a Downed Segment 371 Planning Downtime for Maintenance 371 Bringing

Switch (vDS) Health Check 318 Next Steps 319 Setting Up VMware vSphere DRS and HA 319 Enabling DRS 319 Enabling HA 320 Next Steps 323 Setting Up VMware vSphere Storage 323 Configuring vSAN 323 Validating the Greenplum Installation 394 Managing Single Host Failures in VMware vSphere 395 How DRS and HA protect a running Greenplum cluster 395 Single Host Failure Scenarios 396 Mirror Segment Failure Standby Master Failure 397 Master Failure 398 After Host Recovery 399 Best Practices for Greenplum HA 399 Appendix: Grace Period for a Downed Segment 399 Planning Downtime for Maintenance 400 Bringing

Tanzu Greenplum 6 Documentation VMware, Inc 13 Setting Up VMware vSphere DRS and HA 308 Enabling DRS 309 Enabling HA 310 Next Steps 313 Setting Up VMware vSphere Storage 313 Configuring vSAN 313 Greenplum 6 Documentation VMware, Inc 15 Managing Single Host Failures in VMware vSphere 384 How DRS and HA protect a running Greenplum cluster 384 Single Host Failure Scenarios 385 Mirror Segment Failure Standby Master Failure 387 Master Failure 388 After Host Recovery 388 Best Practices for Greenplum HA 388 Appendix: Grace Period for a Downed Segment 389 Planning Downtime for Maintenance 389 Bringing

Switch (vDS) Health Check 303 Next Steps 304 Setting Up VMware vSphere DRS and HA 304 Enabling DRS 304 Enabling HA 305 Next Steps 308 Setting Up VMware vSphere Storage 308 Configuring vSAN 308 Validating the Greenplum Installation 381 Managing Single Host Failures in VMware vSphere 381 How DRS and HA protect a running Greenplum cluster 381 Single Host Failure Scenarios 382 Mirror Segment Failure Standby Master Failure 383 Master Failure 385 After Host Recovery 385 Best Practices for Greenplum HA 385 Appendix: Grace Period for a Downed Segment 386 Planning Downtime for Maintenance 386 Bringing

Switch (vDS) Health Check 275 Next Steps 276 Setting Up VMware vSphere DRS and HA 276 Enabling DRS 276 Enabling HA 277 Next Steps 279 Setting Up VMware Vsphere Storage 280 Configuring vSAN 280 Validating the Greenplum Installation 322 Managing Single Host Failures in VMware vSphere 323 How DRS and HA protect a running Greenplum cluster 323 Single Host Failure Scenarios 324 Mirror Segment Failure Documentation VMware, Inc 14 Master Failure 326 After Host Recovery 326 Best Practices for Greenplum HA 326 Appendix: Grace Period for a Downed Segment 327 Planning Downtime for Maintenance 327 Bringing

只负责生成和优化查询 计划、派发任务、协调数据节点进行并行计算。 按照我们在用户现场观察到的，Master 上的资源消耗很少有超过 20% 情况发生，因为 Segment 才是计算和加载发生的场所（当然， 在 HA 方面，Greenplum 提供 Standby Master 机制进行保证）。 再进一步看，Master-Slave 架构在业界的大数据分布式计算和云计 算体系中被广泛应用，大家可以看到，现在主流分布式系统都是采 ·网卡绑定采用 mode=4 (802.3ad)，流量传输 hash 策略选用 xmit_hash_policy=layer3+4【((sourceport XOR dest port) XOR ((source IP XOR dest IP) AND 0xffff) modulo )】，这 样能保证流量均匀的打在多块网卡上，另外注意采用 mode4 绑定的 时候，一定要把交换机设置在

Loading Data Using \copy in psql 549 Input Format 549 Running COPY in Single Row Error Isolation Mode 550 Optimizing Data Load and Query Performance 550 Unloading Data from Greenplum Database 551 Defining Database 821 Reloading Configuration File Changes Only 821 Starting the Coordinator in Maintenance Mode 822 Stopping Greenplum Database 822 Stopping Client Processes 823 Managing Greenplum Database Access Input File for System Expansion 912 Creating an input file in Interactive Mode 912 To create an input file in interactive mode 912 Expansion Input File Format 914 Running gpexpand to Initialize New Segments

Greenplum Database 500 Reloading Configuration File Changes Only 500 Starting the Master in Maintenance Mode 500 Stopping Greenplum Database 501 Stopping Client Processes 502 Accessing the Database 502 Accessing Input File for System Expansion 606 Creating an input file in Interactive Mode 606 To create an input file in interactive mode 606 Expansion Input File Format 608 Running gpexpand to Initialize New Segments psql 844 Input Format 845 Running COPY in Single Row Error Isolation Mode 845 Running COPY in Single Row Error Isolation Mode 0 Optimizing Data Load and Query Performance 846 Optimizing Data Load

Greenplum Database 491 Reloading Configuration File Changes Only 491 Starting the Master in Maintenance Mode 491 Stopping Greenplum Database 492 Stopping Client Processes 493 Accessing the Database 493 Accessing Input File for System Expansion 597 Creating an input file in Interactive Mode 597 To create an input file in interactive mode 597 Expansion Input File Format 599 Running gpexpand to Initialize New Segments psql 835 Input Format 836 Running COPY in Single Row Error Isolation Mode 836 Running COPY in Single Row Error Isolation Mode 0 Optimizing Data Load and Query Performance 837 Optimizing Data Load