Adding a drive with data on it to a Concatenation... bad idea?

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

Just as a follow-up to my previous thread on this subject...
I was talking with the tech from the company I bought my
SATA RAID card from, and he claimed that he took an existing
Windows installation partition and added it as a single
concatenated disk using the same RAID controller I bought
from them, and he was able to boot from it with no problems.

Judging by what people have said here, however, taking a drive
with data on it and adding it as a concatenated drive will
definitely do destructive things to the data on the drive.

Is there a definite yes or no on this, or is it really dependent
upon the Concatenaton implementation of the controller?

Incidentally, after all this, I went out and bought one of
the more recent Promise SATA controller cards (non-RAID),
which advertised as being "SATA-II, rev. 2.5 compliant".
Well, that would include hot-plug and port-multiplier
support, according to that spec.

However, I noticed that that card doesn't behave well with
port multipliers... it only sees >1 drive attached that way
if it's attached at boot-time (so port-multiplier when
hot-plugged doesn't work), and they even said they never
even tested it to see if it would WORK with a port-multiplier,
despite their claims that it was up to that spec.

Aren't there ANY hardware manufacturers that make stuff for
the consumer that actually KNOW what their own hardware is
supposed to be able to do?! Bleh!

- Tim

--
 
Spammay Blockay said:
Just as a follow-up to my previous thread on this subject...
I was talking with the tech from the company I bought my
SATA RAID card from, and he claimed that he took an existing
Windows installation partition and added it as a single
concatenated disk using the same RAID controller I bought
from them, and he was able to boot from it with no problems.

Actually, I would have thought that it would recognize an intialized
(ie partitioned) drive automatically without any assistance for com-
patability sake.
Judging by what people have said here, however, taking a drive
with data on it and adding it as a concatenated drive will
definitely do destructive things to the data on the drive.

Oh? I must have missed that. Several people no less you say?
Is there a definite yes or no on this, or is it really dependent
upon the Concatenaton implementation of the controller?

Well, who is to know without buying them all.
Expect that they will use the unused sectors in the first 2 tracks
between MBR and first partition or the last incomplete and unused
cylinder of the drive. There should be plenty of room there to not
interfere with the standard structures.
Incidentally, after all this, I went out and bought one of
the more recent Promise SATA controller cards (non-RAID),
which advertised as being "SATA-II, rev. 2.5 compliant".
Well, that would include hot-plug and port-multiplier
support, according to that spec.

However, I noticed that that card doesn't behave well with
port multipliers... it only sees >1 drive attached that way
if it's attached at boot-time (so port-multiplier when
hot-plugged doesn't work), and they even said they never
even tested it to see if it would WORK with a port-multiplier,
despite their claims that it was up to that spec.

So how is that RAID card doing in that respect?
 
Previously Spammay Blockay said:
Just as a follow-up to my previous thread on this subject...
I was talking with the tech from the company I bought my
SATA RAID card from, and he claimed that he took an existing
Windows installation partition and added it as a single
concatenated disk using the same RAID controller I bought
Judging by what people have said here, however, taking a drive
with data on it and adding it as a concatenated drive will
definitely do destructive things to the data on the drive.

Not neceaasrily. If there is only this single disk in
the array and it is JBOD, it may work.
Is there a definite yes or no on this, or is it really dependent
upon the Concatenaton implementation of the controller?

It is completely dependent on the implementation. The RAID
''definitions'' do not say were to put metadata.
Incidentally, after all this, I went out and bought one of
the more recent Promise SATA controller cards (non-RAID),
which advertised as being "SATA-II, rev. 2.5 compliant".
Well, that would include hot-plug and port-multiplier
support, according to that spec.
However, I noticed that that card doesn't behave well with
port multipliers... it only sees >1 drive attached that way
if it's attached at boot-time (so port-multiplier when
hot-plugged doesn't work), and they even said they never
even tested it to see if it would WORK with a port-multiplier,
despite their claims that it was up to that spec.

Simple. Port multipliers are too new and have implementation
troubles. Wait 2-3 years.
Aren't there ANY hardware manufacturers that make stuff for
the consumer that actually KNOW what their own hardware is
supposed to be able to do?! Bleh!

Wrong question. The problem is that even hardware matures
today at the customer. One of the impacts of ''shareholder
value'', that favours short-term profits.

Arno
 
Previously Spammay Blockay <[email protected]> wrote:

[ snippage ]
Not neceaasrily. If there is only this single disk in
the array and it is JBOD, it may work.

Well, as it turns out, any individual drive attached directly to
the controller, or whatever drive it sees when I connect it to a box
with a port-multiplier in front, that I add as a concatenation
in the controller BIOS, gets added undamaged.

However, for some reason the RAID-management utility that runs under
Windows won't allow you to create a single-drive concatenation (or,
perhaps it might just not let you do it for a single drive on a given
port).

In any case, I couldn't do it. Bleh again. :-/

Also, the latest version of the controller BIOS allows for a
"pass-thru" so that a bunch of drives conntected directly, or,
I would hope, through a port-multiplier, could be seen as non-RAID
drives and used as through a regular non-RAID SATA controller.
But Addonics, in their Great Wisdom, seems to be using an older
version of the BIOS, and doesn't let you upgrade it (cheaper hardware,
I guess, for them). It would've been nice to have used some drives as
a RAID, and others just as regular, non-RAID drives.
It is completely dependent on the implementation. The RAID
''definitions'' do not say were to put metadata.

Thanks -- I was wondering about that.
Simple. Port multipliers are too new and have implementation
troubles. Wait 2-3 years.

So I noticed. *sigh*
This new "multilane SATA" seems to be a way to not have to worry
about using a port-multiplier, from what I can tell. Have you
used that before? Looks just like a bunch of SATA wire-sets
in a single cable.
Wrong question. The problem is that even hardware matures
today at the customer. One of the impacts of ''shareholder
value'', that favours short-term profits.

Hey, a guy can dream, can't he? *dreams of a perfect universe
developers and engineers rule the planet* :-)

- Tim

--
 
[ snippage ] [...]
Simple. Port multipliers are too new and have implementation
troubles. Wait 2-3 years.
So I noticed. *sigh*
This new "multilane SATA" seems to be a way to not have to worry
about using a port-multiplier, from what I can tell. Have you
used that before? Looks just like a bunch of SATA wire-sets
in a single cable.

I have not used them, but as far as I undertsnd it is just
mechanically different fron single lane cables, but electrically
and logically the same.
Hey, a guy can dream, can't he? *dreams of a perfect universe
developers and engineers rule the planet* :-)

As long as you do not plan your real world deployments on those
dreams... ;-)

Arno
 
Back
Top