Обсуждение: postgresql memory problem
Hello Team,
Firstly, I want to thank you all for the amazing help.
I need your guidance with PostgreSQL. So, I have a 3-nodes patroni cluster running PostgreSQL 12.9 version. I got some errors related to a memory issue.
ERROR: could not resize shared memory segment "/PostgreSQL.346141036" to 16776236 bytes: No space left on device
FATAL: out of memory
DETAIL: Failed on request of size 128 in-memory context "CacheMemoryContext"
My PostgreSQL parameters related to memory;
work_mem : 128MB
shared_buffer: 16GB
effective_cache_size: 200GB
maintenance_work_mem: 2GB
My server profile:
free -g
total used free shared buff/cache available
Mem: 377 28 22 0 326 347
Swap: 127 0 127
total used free shared buff/cache available
Mem: 377 28 22 0 326 347
Swap: 127 0 127
df
tmpfs 189G 2.0M 189G 1% /dev/shm
What do you think about this? Can you advise some hints?
Best regards,
KB
Hi Laurenz,
Thank you so much for your suggestion. I set a lower work_mem setting (64MB), then I will monitor for a while.
Best regards,
KB
Laurenz Albe <laurenz.albe@cybertec.at>, 21 Şub 2022 Pzt, 11:23 tarihinde şunu yazdı:
On Mon, 2022-02-21 at 10:56 +0300, Kenny Bachman wrote:
> Firstly, I want to thank you all for the amazing help.
> I need your guidance with PostgreSQL. So, I have a 3-nodes patroni cluster running PostgreSQL 12.9 version. I got some errors related to a memory issue.
>
> ERROR: could not resize shared memory segment "/PostgreSQL.346141036" to 16776236 bytes: No space left on device
>
> FATAL: out of memory
> DETAIL: Failed on request of size 128 in-memory context "CacheMemoryContext"
>
> My PostgreSQL parameters related to memory;
> work_mem : 128MB
> shared_buffer: 16GB
> effective_cache_size: 200GB
> maintenance_work_mem: 2GB
>
> My server profile:
> free -g
> total used free shared buff/cache available
> Mem: 377 28 22 0 326 347
> Swap: 127 0 127
>
> df
> tmpfs 189G 2.0M 189G 1% /dev/shm
>
> What do you think about this? Can you advise some hints?
There should be a memory dump in the PostgreSQL log.
Considering the small size of 128, the problem might be somewhere else.
Yours,
Laurenz Albe
--
Cybertec | https://www.cybertec-postgresql.com