Jasen,
Thanks for this. My last use of ramdisk was ages ago and I've always had the idea that it was just a disk in ram with no capability to spill over to disk.
It appears the mind refuses to acknowledge that this has been the situation many years ago :)
Some google searches returned others asking the same question, surely someone must have properly established this under *nix. I'll keep searching, and post my solution for feedback.
Kind regards
Seref
On Sun, Oct 14, 2012 at 5:54 AM, Jasen Betts
<jasen@xnet.co.nz> wrote:
On 2012-10-10, Seref Arikan <serefarikan@kurumsalteknoloji.com> wrote:
> --f46d0443048225e0e704cbb5e0ee
> Content-Type: text/plain; charset=ISO-8859-1
>
> Thanks Bret,
> I'm concerned about what happens when my functions under high load fills
> the ramdrive with temporary tables I'm using. The advantage of telling
> postgres to use ram with an option to fall back to disk is significantly
> better in terms of uptime.
> However, I was thinking about some mechanism in the middle tier that
> watches the space in the ram drive and redirects queries to functions that
> create temp tables on disk, if ram drive is close to full. That may help me
> accomplish what I'm trying to
That's what operating systems are for, ramdisk is only ever a hint,
is ram is short it will wind up in swap, if ram is plentiful a disk
table will be fully buffered in ram.
--
⚂⚃ 100% natural
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
