Meyer, David R wrote:
> Good Morning,
>
> I have had a question asked of me that I just don't know. Actually, I
> think I know, but I've never tried it.
>
> One of our Open Source products at CA is our Bare Metal Restore product
> for Linux systems. The question is...
>
> "If I havea crash and replace my hard drive, and it happens to be a
> larger hard drive than was in there when the DR media was created, will
> it work?"
It depends. There are geometry issues that can cause confusion. Things
like booting booting may be impacted (different BIOS representations of
geometries of drives causes int13 calls to return incorrect blocks
during bootstrap, particularly in older BIOSes). Geometry changes also
tend to cause fdisk to report incorrect partition tables, and tend to
confuse the admin.
We commonly use larger drives to replace smaller drives in a software
RAID1 mirror-set when replacing older drives that are out of inventory.
The difference here is that the software RAID metadevice "hides" the
BIOS geometry somewhat from the user and bootstrap.
> The limitation is not with our DR option...it just creates an image.
> The question I have is, how will Linux handle it? Will it just see a
> lot more free space, or will bad things happen?
If your image creation takes a snapshot of the drive partitioning, and
restores it correctly when doing the rebuild, Linux will only see the
origional partition sizes (potentially with arcane C/H/S issues in
fdisk, depending on geometry differences due to your BIOS).
With newer BIOSes, LBA32 in lilo/grub, and newer versions of fdisk less
likely to report geometry errors, I think many of these issues go away.
In order to "see" the new space, you can only expand the last partition
on the disk or creat a new partition in that free space. If you are
using a logical volume manager (ala LVM or EVMS), you can add that
physical extent to a disk group and grow your volumes into it.
After adding the space, you will need to use a "growfs" tool like
resize2fs/xfs_growfs/resize_reiserfs to make the space usable by a given
filesystem.
Any good DR image recovery backup tool is going to have to deal with
fixing up any partitioning geometry issues.
Rather than deal with disk imaging systems like Ghost, we backup/restore
using dirvish and rsync - and use an auto-build system to build fresh
systems from bare-metal (including RAID devices), completely unattended.
Purposing a fresh system to a machine in production is trivial as well -
all changes to all systems are maintained as handbuilt packages pushed
to all machines in our enterprise.
Someone please correct me if I'm wrong, but this has been my experience.
Hope this helps.
-- - Ian C. Blenke <icblenke@nks.net> (This message bound by the following: http://www.nks.net/email_disclaimer.html)----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 19:11:51 EDT