PG Bug reporting form <noreply@postgresql.org> writes:
> We are configured postgres 13.2 version in the EKS cluster. We are facing
> high memory consumption. we have 60GB of RAM and with in one week almost
> the memory is reaching 40 GB. I found one of the checkpointer background
> process is consuming 16.1 % memory. I never seen this much memory
> consumption in the postgres. Please provide any option to resolve this
> memory issue. Kindly let me know if any memory leak in the postgres 13.2
> version.
It seems quite unlikely that there's any actual bug here; more likely,
you are misreading the output of "ps". Many versions of ps will report
a process's use of shared memory as though it were local memory in that
process, ie ps will claim that each process is consuming the whole amount
of the shared memory, even though there's really only one copy.
Particularly with the checkpointer, which doesn't run any user-defined
code, a leak isn't very credible. I would ask first what is your
shared_buffers setting? If the checkpointer's reported memory
consumption isn't much more than that, then there's nothing to see here.
regards, tom lane
