Re: Huge Data sets, simple queries - Mailing list pgsql-performance

From PFC
Subject Re: Huge Data sets, simple queries
Date
Msg-id op.s39morhucigqcu@apollo13
Whole thread Raw
In response to Re: Huge Data sets, simple queries  ("Luke Lonergan" <llonergan@greenplum.com>)
Responses Re: Huge Data sets, simple queries
List pgsql-performance
>> Linux does balanced reads on software
>> mirrors.  I'm not sure why you think this can't improve bandwidth.  It
>> does improve streaming bandwidth as long as the platter STR is more than
>> the bus STR.
>
> ... Prove it.
>

    (I have a software RAID1 on this desktop machine)

    It's a lot faster than a single disk for random reads when more than 1
thread hits the disk, because it distributes reads to both disks. Thus,
applications start faster, and the machine is more reactive even when the
disk is thrashing. Cron starting a "updatedb" is less painful. It's cool
for desktop use (and of course it's more reliable).

    However large reads (dd-style) are just the same speed as 1 drive. I
guess you'd need a humongous readahead in order to read from both disks.

pgsql-performance by date:

Previous
From: "Luke Lonergan"
Date:
Subject: Re: Huge Data sets, simple queries
Next
From: "Jim C. Nasby"
Date:
Subject: Re: Huge Data sets, simple queries