PostgreSQL启动报错“could not map anonymous shared memory: Cannot allo ...

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

1. 基础信息
2. OS版本：Red Hat Enterprise Linux Server release 7.9 (Maipo)
3. DB版本：16.2
4. pg软件目录：/home/pg16/soft
5. pg数据目录：/home/pg16/data
6. 端口：5777

复制代码

报错

1. [pg16@test ~]$ pg_ctl start
2. waiting for server to start....2024-06-01 22:20:02.156 PDT [68668] DEBUG:  registering background worker "logical replication launcher"
3. 2024-06-01 22:20:02.157 PDT [68668] DEBUG:  loaded library "pg_stat_statements"
4. 2024-06-01 22:20:02.157 PDT [68668] FATAL:  could not map anonymous shared memory: Cannot allocate memory
5. 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.
6. 2024-06-01 22:20:02.157 PDT [68668] LOG:  database system is shut down
7. stopped waiting
8. pg_ctl: could not start server
9. Examine the log output.

复制代码

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

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

复制代码

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

1. [pg16@test log]$ pg_ctl start
2. waiting for server to start....2024-06-02 06:43:45.043 PDT [3636] DEBUG:  registering background worker "logical replication launcher"
3. 2024-06-02 06:43:45.043 PDT [3636] DEBUG:  loaded library "pg_stat_statements"
4. 2024-06-02 06:43:45.051 PDT [3636] LOG:  redirecting log output to logging collector process
5. 2024-06-02 06:43:45.051 PDT [3636] HINT:  Future log output will appear in directory "log".
6. done
7. server started

复制代码

服膺：心存敬畏，行有所止。

免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！更多信息从访问主页：qidao123.com:ToB企服之家，中国第一个企服评测及商务社交产业平台。