RAID 1 on a server with possible mad mobo

RAID 1 on a server with possible mad mobo

[Colin McDonald]

We have a production server that has apparently crashed and is not
getting to the point of booting.

It is a Fedora Core 3 install using md with a RAID 1 config with two
IDE disks on a Promise Ultra 133 controller. We may have to move it to
a completely different box.

Without taking up to much of y'alls time, what would be the best
solution for moving the RAID array over to a new box?

1. Fresh install on new disk and then pull the data off and startover
with a new RAID 1 config.
2. Try to boot off of the disks after they have been transferred into
the new machine? I know this will cause all kinds of problems with
kernel/devices, etc and probably won't work.

Are there any How-To's, Docs that discuss this. I have googled but
haven't had any joy.

Thanks

cm

Re: RAID 1 on a server with possible mad mobo

[Paul Clements]

Colin McDonald wrote:

> Without taking up to much of y'alls time, what would be the best
> solution for moving the RAID array over to a new box?

> 2. Try to boot off of the disks after they have been transferred into
> the new machine? I know this will cause all kinds of problems with
> kernel/devices, etc and probably won't work.

Actually, I've done this a couple of times with both Red Hat and SUSE 
and it's worked surprisingly well. With the kernel being mostly modular 
and the hardware detection/configuration utilities being pretty advanced 
these days, it's not much of a problem. (I had a minor issue with SUSE 
doing this sort of thing because the MAC address of the NIC was 
different, so the network stuff was not getting configured. On Red Hat 
[and hopefully Fedora is still the same] you should get prompted at 
bootup if there is any hardware to add or remove.)

Especially if you're wanting to keep the system configuration exactly 
the same, this may be the way to go, rather than trying to reconfigure 
everything exactly the way you had it before.

And of course if, after booting the new system with the old disks, you 
find that things are not quite right, you can always re-install at that 
point...

--
Paul

Re: RAID 1 on a server with possible mad mobo

[Colin McDonald]

Thanks for the suggestions.

Let me ask an additional question.

Is it a bad idea to write the grub to a software mirror. Is it written
to a specific disk when this is done?

If I had a corrupt boot loader or boot sector and i needed to rescue.
Would I point to the md device or one of the disks (or both).

TIA


On Tue, 08 Mar 2005 13:48:03 -0500, Paul Clements
<paul.clements@steeleye.com> wrote:
> Colin McDonald wrote:
> 
> > Without taking up to much of y'alls time, what would be the best
> > solution for moving the RAID array over to a new box?
> 
> > 2. Try to boot off of the disks after they have been transferred into
> > the new machine? I know this will cause all kinds of problems with
> > kernel/devices, etc and probably won't work.
> 
> Actually, I've done this a couple of times with both Red Hat and SUSE
> and it's worked surprisingly well. With the kernel being mostly modular
> and the hardware detection/configuration utilities being pretty advanced
> these days, it's not much of a problem. (I had a minor issue with SUSE
> doing this sort of thing because the MAC address of the NIC was
> different, so the network stuff was not getting configured. On Red Hat
> [and hopefully Fedora is still the same] you should get prompted at
> bootup if there is any hardware to add or remove.)
> 
> Especially if you're wanting to keep the system configuration exactly
> the same, this may be the way to go, rather than trying to reconfigure
> everything exactly the way you had it before.
> 
> And of course if, after booting the new system with the old disks, you
> find that things are not quite right, you can always re-install at that
> point...
> 
> --
> Paul
>

Re: RAID 1 on a server with possible mad mobo

[berk walker]

Grub is referenced to your boot drive, then from that it makes all 
references in relation to /.  The dox on that slightly sux, but it is 
really easy to see, if you can change to boot drive in bios. Run it 
crossways, whatever.  The bios disk boots, and will become home of /.  
Well, at least it has for me for 5 yrs.  Now your milage may vary, 
depending upon which disk your are booting, and the various flavours you 
have available on all of your bootable disks.  Xandros is really willing 
to try anything - suse barely likes itself.  The bottom line is to ask 
it.  Set up a configuration and try it.  It is a *good thing* to be 
backed up, tho.... I have seen OSs that automounted file systems rw, and 
then proceeded to trash them, presumably from ignorance.  One thing 
might be - if grub wants hd 0,0 is to have nothing there, I never do.
just a thought.
b-

Colin McDonald wrote:

>Thanks for the suggestions.
>
>Let me ask an additional question.
>
>Is it a bad idea to write the grub to a software mirror. Is it written
>to a specific disk when this is done?
>
>If I had a corrupt boot loader or boot sector and i needed to rescue.
>Would I point to the md device or one of the disks (or both).
>
>TIA
>
>
>On Tue, 08 Mar 2005 13:48:03 -0500, Paul Clements
><paul.clements@steeleye.com> wrote:
>  
>
>>Colin McDonald wrote:
>>
>>    
>>
>>>Without taking up to much of y'alls time, what would be the best
>>>solution for moving the RAID array over to a new box?
>>>      
>>>
>>>2. Try to boot off of the disks after they have been transferred into
>>>the new machine? I know this will cause all kinds of problems with
>>>kernel/devices, etc and probably won't work.
>>>      
>>>
>>Actually, I've done this a couple of times with both Red Hat and SUSE
>>and it's worked surprisingly well. With the kernel being mostly modular
>>and the hardware detection/configuration utilities being pretty advanced
>>these days, it's not much of a problem. (I had a minor issue with SUSE
>>doing this sort of thing because the MAC address of the NIC was
>>different, so the network stuff was not getting configured. On Red Hat
>>[and hopefully Fedora is still the same] you should get prompted at
>>bootup if there is any hardware to add or remove.)
>>
>>Especially if you're wanting to keep the system configuration exactly
>>the same, this may be the way to go, rather than trying to reconfigure
>>everything exactly the way you had it before.
>>
>>And of course if, after booting the new system with the old disks, you
>>find that things are not quite right, you can always re-install at that
>>point...
>>
>>--
>>Paul
>>
>>    
>>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>.
>
>  
>

Re: RAID 1 on a server with possible mad mobo

[Mike Hardy]

Colin McDonald wrote:

  Is it a bad idea to write the grub to a software mirror. Is it written
> to a specific disk when this is done?

The "Software Raid and Grub HOW-TO"

http://lists.us.dell.com/pipermail/linux-poweredge/2003-July/014331.html

I use grub+raid1 on the root drive of a number of machines, but you do 
have to be careful as its my understanding its not raid aware, and 
"picks" a drive, whereas lilo was raid-aware and wrote the boot sector 
on all components.

After following the directions at the link above, I've been able to boot 
the machine off each component (during failure testing), so the 
directions appear to work, to me.

-Mike