Re: BUG #6186: out of memory while analyze - Mailing list pgsql-bugs

From Valentine Gogichashvili
Subject Re: BUG #6186: out of memory while analyze
Date
Msg-id CAP93muV9PTiRFr7gqDTDwFRiS_TJMUb2iq2hS80GgRi7a05CWw@mail.gmail.com
Whole thread Raw
In response to Re: BUG #6186: out of memory while analyze  (Lampa <lampacz@gmail.com>)
Responses Re: BUG #6186: out of memory while analyze  (Lampa <lampacz@gmail.com>)
List pgsql-bugs
>
>
> ssl = false
> shared_buffers = 900MB                  # min 128kB
> work_mem = 100MB                                # min 64kB
> maintenance_work_mem = 524MB            # min 1MB
> fsync = off                             # turns forced synchronization on
> or off
> synchronous_commit = off                # immediate fsync at commit
> full_page_writes = off                  # recover from partial page writes
> effective_cache_size = 1GB
> default_statistics_target = 8000        # range 1-10000
>
>
> Also tried change maintenance_work_mem to 2G and same result.
> Imposible to set over 2GB (FATAL:  3145728 is outside the valid range
> for parameter "maintenance_work_mem" (1024 .. 2097151))
>
>
Actually, it also can be, that your commit limit is reached. Can you also
look in the /proc/meminfo -> CommitLimit and Commit_AS. These values should
not be too close to each other.

What are the values
for /proc/sys/vm/swappiness, /proc/sys/vm/overcommit_memory
and /proc/sys/vm/overcommit_ratio? What is your swap size. They all
influence CommitLimit and the amount of memory, you can get.

-- Valentin

pgsql-bugs by date:

Previous
From: Dave Page
Date:
Subject: Re: BUG #6191: One click installer fails
Next
From: Heikki Linnakangas
Date:
Subject: ECPGdo fails to restore LC_NUMERIC