[CentOS] (no subject) software Raid as backup disk

[dan1]

>Hello,
>
>I would like to replace one of the disks in a raid 1 array (software raid) 
>in centos 4.4 for the purpose of >saving the removed drive as a backup of 
>the system.  Replace it with a new disk and have the raod >resync.  That 
>way the removed disk can be used to restore the system to that point in 
>time if something >dramatic occured.
>
>I have a number of questions, I can’t find the answers to and I don't have 
>a system I can play with to see >how the software behaves:
>
>1) Do I need to partition the replacement drive or will the system do it 
>after reboot?
>
>
>2) Should I break the raid before replacement or just shutdown, replace and 
>reboot?
>
>
>3) I have also read that acronis 10 rescue CD can be booted and take a 
>system image that way but I >have not tried that yet..
>
>In general, are these approaches a good idea for generating an offsite 
>image backup??
>
>Any help or input would be appreciated.
>
>Thank you,
>
>-ed-


Hello, Ed.
We will most probably have to talk together. I have the exact same purpose 
than you have. I am working like this since several years and it is 
absolutely great (with mdadm software raid 1).
I do even go further: I synchronise remotely a complete system on a local 
software raid system with rsync and then I can boot up the backuped system 
whenever I want, and it will just be a working and bootable backup mirror, 
on another system. But it is not always so easy, problems do arise doing 
this, and I am working on it.

To answer your questions:

1) You need to fdisk your partitions yourself on the disk, exactly the same 
way the source partitions have been done, it is not done automatically. Once 
this is done, you need to add the partitions to the working md array. It 
takes time to synchronise everything. All depends on the speed of your disk 
transfer, but if you have about 50Mbytes/sec, you need more than a hour for 
a 200 Gbytes array.

2) You don't need to break the raid (I suppose mark the disk to be removed 
as failed and then remove it from the array), before shutting down. You can 
simply shutdown the system, remove and replace the disk, and then the system 
will start up with only one drive out of two (of course supposing that you 
use raid 1, mirroring).
However, warning: you need to have the boot loader installed on the disk 
which will remain on the system. If this is not the case, there will be no 
reboot at all !
Also, if you are using the boot partion also as raid1 device, you will come 
into troubles, because it won't find theboot partition to boot from. This is 
due to the fact that you define the proper boot partition disk in the 
/boot/grub/devices.map file. This device is different for each drive array, 
and thus it won't work out of the box. Therefore, the proper way to work is 
to have the boot partition not as raid device, but the second partition must 
be a normal ext3 partition, and must be a copy of everything, except the 
/boot/grub/device.map file.

3) No idea, but you should be able to do it through the CentOS 4.4 Live CD 
also.

Regards,
Daniel