hi ya
am curious... why 3Ware cards are not in the list of "hardware raid"
supported chipsets..
- 3ware (supposedly) does raid0/raid1 in hw...
- easy enough to test ... pull the drives... :-)
c ya
alvin
On Fri, 22 Feb 2002, Emil Pedersen wrote:
> O Polite wrote:
> >
> > On Fr
At 03:20 PM 02/22/02 -0800, nate wrote:
>one of my servers is running 3 x seagate ST39204LW
>Ultra160 10,000RPM drives. 2 are in raid0 and 1 is
>not.
>
>this is for a 477MB file.
>
>non raid:
>mail:~# time dd if=/dev/zero of=testfile bs=10 count=5000
>5000+0 records in
>5000+0 records out
>
>re
> On Fri, 2002-02-22 at 02:16, Alvin Oga wrote:
>
>> if you can write data to /dev/hda1 in 5 seconds...
>> writing to striped raid0 will be say 7-10 seconds...
> Why is this? I thoght the point of raid0 was to write to all drives
> in parallell.
it is, and there is a decent performance improvem
O Polite wrote:
>
> On Fri, 2002-02-22 at 02:16, Alvin Oga wrote:
>
> > if you can write data to /dev/hda1 in 5 seconds...
> > writing to striped raid0 will be say 7-10 seconds...
> Why is this? I thoght the point of raid0 was to write to all drives in
> parallell.
It is[1]. You trade increase
On Fri, 2002-02-22 at 02:16, Alvin Oga wrote:
> if you can write data to /dev/hda1 in 5 seconds...
> writing to striped raid0 will be say 7-10 seconds...
Why is this? I thoght the point of raid0 was to write to all drives in
parallell.
op
* Alvin Oga ([EMAIL PROTECTED]) spake thusly:
>
> ji ya dimi
>
> >
> > There are motherboards with hardware ATA-100 RAID controllers,
> > these would give you the best bang per buck (if the controller
> > is well supported by the kernel): cheap IDE drives + high
> > performance of RAID.
>
> non
hi ya ron
yuppers... on raid1 mirror and it must die for raid0...
- problem with onboard raid1(mirroring)
- if one removes the disks... it still dies...
and to test if hw stripping works ( raid0 )... lot tougher...
if you can write data to /dev/hda1 in 5 seconds...
writing to s
On Thu, 21 Feb 2002 16:11:33 -0800 (PST) Alvin Oga <[EMAIL PROTECTED]> wrote:
[snip]
>
> none of those onboard raid stuff works ...
>
> http://www.linux-ide.org/chipsets.html
>
>
> true test is...
> - install your root-raid setup
> - reboot make sure it normally all works
ji ya dimi
>
> There are motherboards with hardware ATA-100 RAID controllers,
> these would give you the best bang per buck (if the controller
> is well supported by the kernel): cheap IDE drives + high
> performance of RAID.
none of those onboard raid stuff works ...
http://www.linux-
on Thu, Feb 21, 2002, nate ([EMAIL PROTECTED]) wrote:
>
>
> > First I think that I should get faster disk I/O than I have today.
> > So for the first time ever I'm considering some RAID solution.
> >
> > In your experience what will get the best price / speed performance
> > ratio?
> >
> > Softwa
> First I think that I should get faster disk I/O than I have today.
> So for the first time ever I'm considering some RAID solution.
>
> In your experience what will get the best price / speed performance
> ratio?
>
> Software IDE/RAID
> Software SCSI/RAID
> IDE RAID controller card
> SCSI RAID
* O Polite ([EMAIL PROTECTED]) spake thusly:
> I'll be building a new development machine soon. It's been a while since
> the last time, and it's seems to be a lot harder today than the last
> time I did it a few years ago. There are so many CPU slots and memory
> types to choose from.
>
> First I
hi ya "o polite"
raid0(striping) or raid5 is faster on reads... and generally slower
on writes... 2x or 5x more disks to write into... more
additional parity calculations
it'd help if you can specify your budget and/or disk capacity
requirements or "purpose of the raid" system...
raid i
I'll be building a new development machine soon. It's been a while since
the last time, and it's seems to be a lot harder today than the last
time I did it a few years ago. There are so many CPU slots and memory
types to choose from.
First I think that I should get faster disk I/O than I have toda
14 matches
Mail list logo
