Hi,
I'm using the SystemImager 4.0.2 stable release on an HP ProLiant DL380
G5 server with CentOS 5.4. It's a 32 bit OS on an i686 architecture.
I've noticed an issue where disk partitions get 5% smaller each time I
do a system restore because of an incompatibility in the units that are
used during the image and during the restore.
I dug into the code a little bit and found that the partition sizes are
recorded as MB = 2^20 Bytes = 1,048,576 bytes during the image using a
command like "sfdisk -l -uM /dev/mydiskname". Then those sizes are
passed to "parted" during the restore when creating the partitions.
The issue is that "parted" units are in MB = 1,000,000 bytes. So the
result is that all partitions except the last one on the disk shrink
by 5% and the final partition is assigned the remaining extra space.
My solution is to comment out lines 957, 958 and 959 of
"/usr/sbin/si_prepareclient". ie: To comment out these lines:
if($arch eq "i386") {
$preferred_tool = 'sfdisk';
}
This forces SystemImager to use "parted" when reading disk sizes during
imaging on the x86 architecture. I can see a difference in the
partition sizes listed in "/etc/systemimager/autoinstallscript.conf"
before and after making the change listed above. So, there seems to be
a difference in the partition sizes when they are read using parted and
when they're read using sfdisk.
Has anyone else seen this issue? Is there a recommended workaround?
Thanks,
Tyler
------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
sisuite-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sisuite-users
