FAI disk config problem

[RU Admin]

Thank you for your response.  It seems as though the align-at was the 
problem.   It appears that if you don't specify "align-at", the checks 
that are done with "preserve_always" don't get the "default" value of 
"align-at", but if you specifically set it, then it's fine.

So in other words, I simply modified my disk_config and added 
"align-at:1M", then did a fresh FAI on the system (without the 
"preserve_always" option), and again the first partition was offset by 1M 
(just like I had before).  But this time, when did the FAI again with the 
"preserve_always" option, I didn't get any "Previous partitions 
overflow..." error and setup-storage properly preserved the partitions 
properly.  So from now on, I'll just add "align-at:1M" to all of my 
disk_config's, however this does seem to be a little bug, although not 
critical since there is an easy work-around.

Craig

On Wed, 10 Oct 2012, Sylvain Milot wrote:

>
> Hello,
>
> sorry for top-posting ...
>
> I would guess that the "align-at" feature is set to 1M (2048 sectors)
> by default for this setup-storage version ? I can't say for sure as I
> don't use this version ... still using 3.4.8, but I always set align-at:1M.
>
> Maybe you'd like to read this document if unfamiliar:
> http://www.ibm.com/developerworks/linux/library/l-4kb-sector-disks/
>
> or this even:
> http://www.novell.com/support/kb/doc.php?id=7007193
>
> You could possibly use align-at:1K to force it, if that's really what
> you want ?
>
> I'll let others answer your other question.
>
> Best,
>
> Sylvain
>
> On Wed, 10 Oct 2012, RU Admin wrote:
>
>> 
>> I'm running Debian Wheezy with FAI 4.0.3 and for my disk_config I'm using 
>> the following:
>> 
>> disk_config disk1 disklabel:msdos fstabkey:uuid bootable:1
>> primary  /            20G      ext4  rw,noatime,errors=remount-ro 
>> tuneopts="-c 0 -i 0"
>> logical  /var         10G      ext4  rw,noatime createopts="-L var -m 5" 
>> tuneopts="-c 0 -i 0"
>> logical  swap         8G       swap  sw
>> logical  /tmp         20G      ext4  rw,noatime createopts="-L tmp -m 0" 
>> tuneopts="-c 0 -i 0"
>> logical  /home        1G       ext4  rw,noatime,nosuid createopts="-L home 
>> -m 1"    tuneopts="-c 0 -i 0"
>> logical  /backup      10G-     ext4  rw,noatime createopts="-L backup" 
>> tuneopts="-c 0 -i 0"
>> 
>> My problem is that after I do the FAI install on the machine, and look at 
>> the partition table with fdisk, I see:
>> 
>> /dev/sda1   *        2048    41945087    20971520   83  Linux
>> /dev/sda2        41945088  1952448511   955251712    f  W95 Ext'd (LBA)
>> /dev/sda5        41945090    62916609    10485760   83  Linux
>> /dev/sda6        62916612    79693827     8388608   82  Linux swap / 
>> Solaris
>> /dev/sda7        79693830   121636869    20971520   83  Linux
>> /dev/sda8       121636872   123734023     1048576   83  Linux
>> /dev/sda9       123734026  1952448511   914357243   83  Linux
>> 
>> Why is the first partition starting at 2048 instead of 1?  If I look at the 
>> disk in cfdisk, it's showing 1.05MB of free space at the beginning of the 
>> disk...although it's a trivial amount of space being lost, why am I losing 
>> it?
>> 
>> I have a Debian Squeeze system running FAI 3.4.7 and I don't see that 
>> problem when using the same disk config.
>> 
>> Any help would be much appreciated, the main reason I'm looking into this 
>> is because I'm trying to use the preserve_always option with FAI 4.0.3 
>> after I've already done an initial install on the system with the same FAI 
>> disk_config and I'm getting:
>> 
>> Previous partitions overflow begin of preserved partition /dev/sda8
>> Prototype mismatch: sub Parse::RecDescent::namespace000001::dclone ($) vs 
>> none at /usr/share/perl5/Parse/RecDescent. pm line 1984 during global 
>> destruction.
>> 
>> I was hoping it had something to do with the 1.05MB of wasted space in the 
>> beginning of the disk causing problems.  Again, using the same disk_config 
>> and adding the preserve_always works perfectly with 3.4.7.
>> 
>> Thanks!
>> 
>> Craig
>> 
>