PostgreSQL启动报错“could not map anonymous shared memory: Cannot allocate memory”

PostgreSQL启动报错"could not map anonymous shared memory: Cannot allocate memory"

bash 复制代码
基础信息
OS版本:Red Hat Enterprise Linux Server release 7.9 (Maipo)
DB版本:16.2
pg软件目录:/home/pg16/soft
pg数据目录:/home/pg16/data
端口:5777

报错

bash 复制代码
[pg16@test ~]$ pg_ctl start
waiting for server to start....2024-06-01 22:20:02.156 PDT [68668] DEBUG:  registering background worker "logical replication launcher"
2024-06-01 22:20:02.157 PDT [68668] DEBUG:  loaded library "pg_stat_statements"
2024-06-01 22:20:02.157 PDT [68668] FATAL:  could not map anonymous shared memory: Cannot allocate memory
2024-06-01 22:20:02.157 PDT [68668] HINT:  This error usually means that PostgreSQL's request for a shared memory segment exceeded available memory, swap space, or huge pages. To reduce the request size (currently 295698432 bytes), reduce PostgreSQL's shared memory usage, perhaps by reducing shared_buffers or max_connections.
2024-06-01 22:20:02.157 PDT [68668] LOG:  database system is shut down
 stopped waiting
pg_ctl: could not start server
Examine the log output.

PostgreSQL 请求的共享内存段超出了当前可用的内存、交换空间或大页内存 (huge pages)。当前请求的内存大小为 295,698,432 字节(约 282 MB)。

分析记录

bash 复制代码
[pg16@test ~]$ cat data/postgresql.conf |grep shared
shared_buffers = 256MB                  # min 128kB
#shared_memory_type = mmap              # the default is the first option
dynamic_shared_memory_type = posix      # the default is usually the first option
#min_dynamic_shared_memory = 0MB        # (change requires restart)
#wal_buffers = -1                       # min 32kB, -1 sets based on shared_buffers
shared_preload_libraries = 'pg_stat_kcache,pg_stat_statements,auto_explain'     # (change requires restart)
[pg16@test ~]$ cat /proc/meminfo |grep Huge
AnonHugePages:    942080 kB
HugePages_Total:     150
HugePages_Free:        7
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB

发现空闲的大页只有7个,不满足pg的启动需求,故报错。最终发现是其它数据库进程占用了,停了之后pg可以正常启动。

解决

bash 复制代码
[pg16@test log]$ pg_ctl start
waiting for server to start....2024-06-02 06:43:45.043 PDT [3636] DEBUG:  registering background worker "logical replication launcher"
2024-06-02 06:43:45.043 PDT [3636] DEBUG:  loaded library "pg_stat_statements"
2024-06-02 06:43:45.051 PDT [3636] LOG:  redirecting log output to logging collector process
2024-06-02 06:43:45.051 PDT [3636] HINT:  Future log output will appear in directory "log".
 done
server started

谨记:心存敬畏,行有所止。

相关推荐
谙弆悕博士40 分钟前
多智能体架构详解:层级树状架构(分级主从型)
运维·架构·大模型·agent·多智能体·langgraph·层级树状
bksczm2 小时前
Redis Cluster 分片集群详解:槽路由、故障判定与扩缩容
数据库·redis·分布式
HMS工业网络3 小时前
工业远程访问的五项安全部署原则
运维·服务器·安全
黑十3 小时前
FDE 怎么快速搭一个可演示页面?最小前端方案
java·大数据·前端·数据库·人工智能
Databend4 小时前
OpenRaft 自定义 Membership,按部署拓扑选择 Quorum
大数据·数据库·数据分析
Databend4 小时前
用 SQL 排查支付故障,OpenTelemetry 接入 Databend Cloud 实战
大数据·数据库·sql
吴建旭 智宅焕4 小时前
智能家居全国交付中执行端能力评估与分级调度架构设计:从注册数量到交付确定性的系统实现
运维·智能家居·devops
kyle~4 小时前
SQLite -- 低成本持久化能力
数据库·sqlite
上海潭龙东海科技有限公司5 小时前
智慧堆场动态模型三维实时孪生重构智能管控
大数据·网络·数据库·算法·重构
老纪的技术唠嗑局5 小时前
OceanBase AI数据大会,我会这么逛
数据库