gp版本：6.21.2

报错内容："LOG","00000","could not fork new process for connection: Cannot allocate memory",,,,,,,0,,"postmaster.c",4532,

问题描述：持续不到2分钟造成节点宕机。

根据监控系统最低剩余60GB可用内存，应该是GP后台进程占用内存超过参数限制导致，应该调整哪些参数比较合适？

每个数据节点物理内存250GB,6个primary segment,6个mirror segment。下面是相关参数。

Values on all segments are consistent
GUC : gp_vmem_protect_limit
Master value: 39100
Segment value: 39100

Values on all segments are consistent
GUC : max_statement_mem
Master value: 2GB
Segment value: 2GB

Values on all segments are consistent

GUC : shared_buffers

Master value: 8GB

Segment value: 8GB

Values on all segments are consistent
GUC : statement_mem
Master value: 256MB
Segment value: 256MB

Values on all segments are consistent
GUC : gp_resource_manager
Master value: queue
Segment value: queue

Values on all segments are consistent
GUC : max_resource_queues
Master value: 9
Segment value: 9

Values on all segments are consistent
GUC : gp_enable_global_deadlock_detector
Master value: on
Segment value: off

Values on all segments are consistent
GUC : work_mem
Master value: 32MB
Segment value: 32MB

Values on all segments are consistent
GUC : max_connections
Master value: 800
Segment value: 2000