Re: FreeSpaceMap hashtable out of memory

It is problematic to produce small enough subset, due to large DB and
randomness of the situation.
But here is what we see in server log file, see below:

It seems like
WARNING: ShmemAlloc: out of memory
ERROR: FreeSpaceMap hashtable out of memory

goes together, does it related to the size of Shared Memory or could
increase of
Shared Memory solve the thing or that just a coincidence?

Another thing ( from the code) it seems that
hash_search trying to inset new entry into DynHash and cannot allocate
memory
in DynHash Context, so another question is DynHash Context upper
limited, another word
cannot grow more than 8 * 1024 * 1024 byte?

Any suggestions are welcome,
Thank you.

------------------- LOG
----------------------------------------------------
LOG: all server processes terminated; reinitializing shared memory and
semaphores
IpcMemoryCreate: shmget(key=5432001, size=4669440, 03600) failed: Not
enough
core

This error usually means that PostgreSQL's request for a shared
memory segment exceeded available memory or swap space.
To reduce the request size (currently 4669440 bytes), reduce
PostgreSQL's shared_buffers parameter (currently 256) and/or
its max_connections parameter (currently 128).

The PostgreSQL Administrator's Guide contains more information about
shared memory configuration.

Error 2:
WARNING: ShmemAlloc: out of memory
ERROR: FreeSpaceMap hashtable out of memory

-----Original Message-----
From: Tom Lane [mailto:tgl(at)sss(dot)pgh(dot)pa(dot)us]
Sent: Wednesday, October 01, 2003 2:51 PM
To: Maksim Likharev
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: [HACKERS] FreeSpaceMap hashtable out of memory

"Maksim Likharev" <mlikharev(at)aurigin(dot)com> writes:
> Using PG under Cygwin we having following error message during INSERT
> INTO
> "FreeSpaceMap hashtable out of memory".

Hm, that's not supposed to happen. Can you create a reproducible
example?

regards, tom lane