HowToFixComputers.com




Watched TopicsWatched Topics SearchSearch RegisterRegister Log in to check your private messagesLog in to check your private messages ProfileProfile Log inLog in
can I move these old disks to new server?

 
Post new topic   Reply to topic    Index -> SCSI
Author Message
SherryB
Guest





PostPosted: Thu Feb 22, 2007 9:59 pm    Post subject: can I move these old disks to new server? Reply with quote

I have an old Gateway 7250R server with hot-swappable (non-RAID) U160
SCSI disk interface. I want to move my 2 newer larger disks to my
newer Gateway 975 Server with hot-swappable (non-RAID) U320 SCSI
disks. The old server is running Windows 2000 Server, and the new one
is running Windows 2003 Server. All the disks are formatted NTFS, and
my goal is to do this without losing the data.
The 2 disks in question are:
1. 73GB IBM Ultrastar 73LZX U160 pn#07N6310
2. 147GB Maxtor Atlas 10K rpm U320 SCA-2 pn#8B146J0

I am pretty sure #2 will work as long as there are no O/S issues (are
there?). But what about putting a U160 disk into a U320 backplane?
Will that slow down all the other disks? Will it even work?

Is there any difference between SCA and SCA-2 interfaces? My old
server says SCA, and new one says SCA-2. (But the info with the Atlas
drive says its SCA-2 and its working in an SCA system now.)

Thanks!
Back to top
Folkert Rienstra
Guest





PostPosted: Sat Mar 03, 2007 4:01 pm    Post subject: Re: can I move these old disks to new server? Reply with quote

"Michael Baeuerle" <michael.baeuerle@gmx.net> wrote in message news:4cblb4-ku.ln1@micha.freeshell.org
Quote:
Folkert Rienstra wrote:

Michael Baeuerle wrote:

The 73GByte U160 IBM drive is not old enough.

Huh?

It must be SCA-2 compliant even if it is labeled "SCA".

And this has what to do with SCA and SCA-2 not "be the same" exactly?

As you stated earlier their is no "SCA" defined by the SCSI standard -
so nobody knows what "SCA" exactly is.

I have written that both can be considered the same (I mean
_for_this_case_ - plugging U160 drives in an U320 backplane)
because all U160 drives labeled "SCA" simply have the wrong term
printed on them but are in fact SCA-2 and things should always work.

OP's previous system was SCA(-1). OP's new one is SCA-2.
OP never mentioned that the IBMs were SCA. You made that up.
You are just babbling.

Quote:

But even if SCA and SCA-2 are not "the same" they are compatible
enough (as I expect it from SCSI) that you can mix SCA-2 and older
SCSI2 "SCA" drives. I have such a configuration running and it works.

So has OP. This wasn't disputed at all.

Quote:

If you think that hotplug may be a problem in this context:
Hotplug events with the bus in reset state should be no problem
with selfmade SCA, SCA-2 or mixed configurations.

Not even with 68-pin if you are normally cautious.
But you may end up with running in asynchronous mode.

Quote:
The higher levels should IMHO only be used on systems where all
components are certified to support them in the used combination.

Including the OS. And if you can get the bus in reset state on a
running system you are already in that higher levels category.

Else you would be holding the reset switch, which is silly.
That's one step removed from switching the system off and just do
the swap cold.

Quote:


Micha
Back to top
Michael Baeuerle
Guest





PostPosted: Sun Mar 04, 2007 6:04 pm    Post subject: Re: can I move these old disks to new server? Reply with quote

Folkert Rienstra wrote:
Quote:

Michael Baeuerle wrote:

[...]
I have written that both can be considered the same (I mean
_for_this_case_ - plugging U160 drives in an U320 backplane)
because all U160 drives labeled "SCA" simply have the wrong term
printed on them but are in fact SCA-2 and things should always work.

OP's previous system was SCA(-1). OP's new one is SCA-2.
OP never mentioned that the IBMs were SCA. You made that up.
You are just babbling.

Hmm, OP wrote:
| My old server says SCA, and new one says SCA-2. (But the info with
| the Atlas drive says its SCA-2 and its working in an SCA system now.)

I read this as "both servers have SCA backplanes" and the 2 disks that
should be moved are both SCA too (because they currently run in the old
"SCA" backplane).

Quote:
[...]
If you think that hotplug may be a problem in this context:
Hotplug events with the bus in reset state should be no problem
with selfmade SCA, SCA-2 or mixed configurations.

Not even with 68-pin if you are normally cautious.
But you may end up with running in asynchronous mode.

Only if the hostadapter driver is crippled. Because any device on the
bus is allowed to create a bus reset, the hostadapter must notice
external bus resets. If it do so, it knows when it's time to
renegotiate.

Quote:
The higher levels should IMHO only be used on systems where all
components are certified to support them in the used combination.

Including the OS.

Ack.

Quote:
And if you can get the bus in reset state on a
running system you are already in that higher levels category.

Else you would be holding the reset switch, which is silly.
That's one step removed from switching the system off and just do
the swap cold.

Ack, but reseting the whole system should only be required if one of the
disks that hold the OS itself must be replaced. Otherwise a switch that
simply pulls down the RST line on the SCSI bus should be sufficient.
With a bistable switch you have both hands free. Another option is to
create the bus reset with the hostadapter (if the OS supports
user-defined timings).

With Unix I would try it like this:
It is obvious that it is not allowed to reset the disk(s) which hold the
root filesystem of the running OS. Therefore I would connect the OS
disk(s) to a unique SCSI bus without hotplug support. A second SCSI bus
is dedicated for the hot-pluggable data drives. If one of them must be
replaced, all filesystems on this bus must be unmounted. The motor of
the disk that should be removed can be stopped now. It is now allowed to
switch the bus to reset state and replace the drive. After spinning up
the drive and reenabling the bus, the OS will renegotiate with all disks
automatically (because it has created or detected the bus reset). After
rebuilding potential RAID arrays and remounting the filesystems, things
are working again. A script can be used to do all this tasks with two
admin commands.


Micha
--
HTML mails are rejected
Back to top
Display posts from previous:   
Post new topic   Reply to topic    Index -> SCSI All times are GMT
Page 1 of 1

 

 MemberlistMemberlist  UsergroupsUsergroups



Powered by p|-|pBB

Featured Sites: DIY Projects