From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
---|---|
To: | Dave Page <dpage(at)pgadmin(dot)org> |
Cc: | Gurjeet Singh <singh(dot)gurjeet(at)gmail(dot)com>, Lizzy M <lizzymy(at)gmail(dot)com>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Can postgresql store its data on raw device now? |
Date: | 2009-06-09 14:12:37 |
Message-ID: | b42b73150906090712le2873a9gde8044f72536ca39@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Tue, Jun 9, 2009 at 4:14 AM, Dave Page<dpage(at)pgadmin(dot)org> wrote:
> On Tue, Jun 9, 2009 at 8:49 AM, Gurjeet Singh<singh(dot)gurjeet(at)gmail(dot)com> wrote:
>> It still doesn't support raw device.
>
> Nor is it ever likely to. Filesystems are vastly superior now than
> when certain other DBMSs were designed to use raw devices, and for
> PostgreSQL to replace the average filesystem now with something as
> reliable, performant and well tested would take a significant amount
> of effort for little, if any, gain.
As the 'solid state drive' train continues to leave the station, this
will become more and more true. Flash based SSD and hard drives have
very different operating characteristics that require cooperation
between the device, filesystem, and o/s if you want things to work
optimally.
Why reinvent the wheel? I think the trend is in in the other
direction...applications are trying to dispatch as much work in
dealing with I/O into the kernel and out of user space as possible.
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | Merlin Moncure | 2009-06-09 14:34:27 | Re: Why lots of temp schemas are being created |
Previous Message | Grzegorz Jaśkiewicz | 2009-06-09 14:11:51 | Re: postgres getting slow under heavy load though autivacuum is enabled |