On Mon, May 17, 2004 at 12:36:24AM -0700, Karsten M. Self wrote:
> This is really old, but it's straight up my alley, so...
>
> on Sat, Apr 03, 2004 at 07:08:39PM -0600, Christopher L. Everett ([EMAIL PROTECTED])
> wrote:
> > I do a lot of database work. Sometimes I must do massive batch jobs on
This is really old, but it's straight up my alley, so...
on Sat, Apr 03, 2004 at 07:08:39PM -0600, Christopher L. Everett ([EMAIL PROTECTED])
wrote:
> I do a lot of database work. Sometimes I must do massive batch jobs on
> my box
> such as:
>
> -- multi-gigabyte database dumps and restores
Glenn Meehan wrote:
On Sun, 2004-04-04 at 11:51, Christopher L. Everett wrote:
So, what kind of hard drive subsystem can I run that would get me
3 to 4 times the performance that wouldn't break the bank?
Which file system are you using?
I find the performance of ext2 superior to that of ext
At 2004-04-04T14:54:08Z, Chris Metzler <[EMAIL PROTECTED]> writes:
> On Sat, 03 Apr 2004 23:42:27 -0600
> Kirk Strauser <[EMAIL PROTECTED]> wrote:
>> Not really. Concurrent mirrored reads can be significantly faster than
>> striped reads, since each drive can be servicing a complete read request
On Sat, 03 Apr 2004 21:28:16 -0600
"Christopher L. Everett" <[EMAIL PROTECTED]> wrote:
>
> I just read an email of the Linux kernel list saying that Linux software
>
> RAID
> kicks the ass of most hardware raid solutions.
Do you have a pointer to that email? Does that email include
benchmarking
On Sat, 03 Apr 2004 23:42:27 -0600
Kirk Strauser <[EMAIL PROTECTED]> wrote:
> At 2004-04-04T05:20:43Z, "Christopher L. Everett"
> <[EMAIL PROTECTED]> writes:
>
> > substitute raid 0 (striping) for raid 1, and that last makes sense.
>
> Not really. Concurrent mirrored reads can be significantly f
On Sun, 2004-04-04 at 11:51, Christopher L. Everett wrote:
> So, what kind of hard drive subsystem can I run that would get me
> 3 to 4 times the performance that wouldn't break the bank?
Which file system are you using?
I find the performance of ext2 superior to that of ext3.
Glenn
--
To U
First thing is I have no experience at all with SCSI, so I'll only talk
about IDE.
Hardware raid is 'easier' as you don't have to load any driver in
booting process. Anyway, there are few pure IDE hardware raids (most
need a driver to properly work) and are a bit more expensive than the
others. Wi
At 2004-04-04T05:20:43Z, "Christopher L. Everett" <[EMAIL PROTECTED]> writes:
> substitute raid 0 (striping) for raid 1, and that last makes sense.
Not really. Concurrent mirrored reads can be significantly faster than
striped reads, since each drive can be servicing a complete read request
simu
Christopher L. Everett wrote:
>>1) Dedicate a second hard drive to the database files
> I'm not sure that would be a huge win, since the performance would just
> be that of another IDE drive.
No, not huge, but it does mean that the rest of the system won't be stealing
disk I/O from your database
Christopher L. Everett wrote:
IIRC, RAID 1 gives the best performance under all conditions :),
though it's
not terribly safe.
substitute raid 0 (striping) for raid 1, and that last makes sense.
--
Christopher L. Everett
Chief Technology Officer www.medbanner.com
Me
Adam Aube wrote:
Christopher L. Everett wrote:
So, what kind of hard drive subsystem can I run that would get me
3 to 4 times the performance that wouldn't break the bank?
Define "break the bank". The more money you're willing to sink into this,
the better performance you can get. A few opti
Christopher L. Everett wrote:
>>>Am I correct in thinking that the bottleneck lies in the HD subsystem?
>>Yes.
> So, what kind of hard drive subsystem can I run that would get me
> 3 to 4 times the performance that wouldn't break the bank?
Define "break the bank". The more money you're willing
Adam Aube wrote:
Christopher L. Everett wrote:
The problem is that these things often take 10 to 30 minutes to run on
my box. When I use the GNU time utility, I see a low PCPU number,
typically between 15 and 25%. CPU utilization viewed through top remains
at 35% or so, and I never go deeper
Kent West wrote:
Christopher L. Everett wrote:
Hi,
I do a lot of database work. Sometimes I must do massive batch jobs
on my box
such as:
-- multi-gigabyte database dumps and restores
-- tests over millions of records, searching for overlooked cases
-- one-off queries for sales & marketing
Christopher L. Everett wrote:
> I do a lot of database work. Sometimes I must do massive batch jobs on
> my box
> The problem is that these things often take 10 to 30 minutes to run on
> my box. When I use the GNU time utility, I see a low PCPU number,
> typically between 15 and 25%. CPU utiliz
Christopher L. Everett wrote:
Hi,
I do a lot of database work. Sometimes I must do massive batch jobs
on my box
such as:
-- multi-gigabyte database dumps and restores
-- tests over millions of records, searching for overlooked cases
-- one-off queries for sales & marketing typs that join 8
17 matches
Mail list logo
