Posts made by Eric Johnson

Ok, testing

[code]
— fog.upload.dist 2014-06-19 09:07:10.000000000 -0400
+++ fog.upload 2014-06-19 09:32:02.000000000 -0400
@@ -94,8 +94,8 @@
fi
fi
diskSize=parted -s $hd u kB print | grep "Disk" | grep -v "Flags" | sed 's/Disk \/dev\/[a-zA-Z0-9]*: //' | sed 's/kB//';

•                   partitionStart=`fdisk -lu | grep $part | awk '{print $3}'`;
                    if [ "$osid" == "1" ]; then
  

•                           partitionStart=`fdisk -lu | grep $part | awk '{print $3}'`;
                            if [ "$partitionStart" == "" ]; then
                                    partitionStart="63";
                            fi
  

@@ -136,22 +136,22 @@
echo “Done”;
if [ “$osid” == “5” ] || [ “$osid” == “6” ] || [ “$osid” == “7” ]; then
if [ “$win7partcnt” == “1” ]; then

•                                                   win7part1start=`parted -s $hd u kB print | sed -e '/^.1/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//'`;
  

•                                                   win7part1start=`parted -s $hd u kB print | sed -e '/^.1/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//' -e 's/\..*$//'`;
                                                    if [ "$win7part1start" == "" ]; then
                                                            handleError "Unable to determine disk start location.";
                                                    fi
                                                    adjustedfdsize=`expr $sizefd '+' $win7part1start`;
                                                    parted -s $hd u kB rm 1 &>/dev/null;
  

•                                                   parted -s $hd -a opt u kB mkpart primary ntfs 2048s ${adjustedfdsize}kB &>/dev/null;
  

•                                           parted -s $hd u kB set 1 boot on &>/dev/null
  

•                                                   parted -s $hd -a opt u kB mkpart primary ntfs ${partitionStart}s ${adjustedfdsize}kB &>/dev/null;
  

•                                                   parted -s $hd u kB set 1 boot on &>/dev/null
                                            elif [ "$win7partcnt" == "2" ]; then
  

•                                                   win7part2start=`parted -s $hd u B print | sed -e '/^.2/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//'`;
  

•                                                   win7part2start=`parted -s $hd u B print | sed -e '/^.2/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//' -e 's/\..*$//'`;
                                                    if [ "$win7part2start" == "" ]; then
                                                            handleError "Unable to determine disk start location.";
                                                    fi
                                                    adjustedfdsize=`expr $sizefd '+' $win7part2start`;
                                                    parted -s $hd -a opt u kB mkpart primary ntfs ${defaultpart2start}B ${adjustedfdsize}kB &>/dev/null;
  

•                                           parted -s $hd u kB set 1 boot on &>/dev/null
  

•                                                   parted -s $hd u kB set 1 boot on &>/dev/null
                                            else
                                                    handleError "Invalid partition count.";
                                            fi
  

[/code]

Well… that diff post didn’t work too well eh? Ill try a file upload…

[url=“/_imported_xf_attachments/1/1022_diff.txt?:”]diff.txt[/url]

Yep, it is a IDE disk in a VirtualBox vm.

So I am wondering why not just use the partitionStart variable when having parted resize the disk. IE make no assumptions about what/where the single partition starts. (not sure about multi-partition setups, we don’t use them here)

Here is a diff for fog.upload that fixes this and the decimal point problem mentioned earlier. When doing the steps “manually” this seems to work. Ill be trying a full script run upload later this morning.

(don’t know if this is a good place to send bug fix diffs, so enlighten me if there is a better way)

E

— fog.upload.dist 2014-06-19 09:07:10.000000000 -0400
+++ fog.upload 2014-06-19 09:32:02.000000000 -0400
@@ -94,8 +94,8 @@
fi
fi
diskSize=parted -s $hd u kB print | grep "Disk" | grep -v "Flags" | sed 's/Disk \/dev\/[a-zA-Z0-9]*: //' | sed 's/kB//';

•                   partitionStart=`fdisk -lu | grep $part | awk '{print $3}'`;
                    if [ "$osid" == "1" ]; then
  

•                           partitionStart=`fdisk -lu | grep $part | awk '{print $3}'`;
                            if [ "$partitionStart" == "" ]; then
                                    partitionStart="63";
                            fi
  

@@ -136,22 +136,22 @@
echo “Done”;
if [ “$osid” == “5” ] || [ “$osid” == “6” ] || [ “$osid” == “7” ]; then
if [ “$win7partcnt” == “1” ]; then

•                                                   win7part1start=`parted -s $hd u kB print | sed -e '/^.1/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//'`;
  

•                                                   win7part1start=`parted -s $hd u kB print | sed -e '/^.1/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//' -e 's/\..*$//'`;
                                                    if [ "$win7part1start" == "" ]; then
                                                            handleError "Unable to determine disk start location.";
                                                    fi
                                                    adjustedfdsize=`expr $sizefd '+' $win7part1start`;
                                                    parted -s $hd u kB rm 1 &>/dev/null;
  

•                                                   parted -s $hd -a opt u kB mkpart primary ntfs 2048s ${adjustedfdsize}kB &>/dev/null;
  

•                                           parted -s $hd u kB set 1 boot on &>/dev/null
  

•                                                   parted -s $hd -a opt u kB mkpart primary ntfs ${partitionStart}s ${adjustedfdsize}kB &>/dev/null;
  

•                                                   parted -s $hd u kB set 1 boot on &>/dev/null
                                            elif [ "$win7partcnt" == "2" ]; then
  

•                                                   win7part2start=`parted -s $hd u B print | sed -e '/^.2/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//'`;
  

•                                                   win7part2start=`parted -s $hd u B print | sed -e '/^.2/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB  .*//' -e 's/\..*$//'`;
                                                    if [ "$win7part2start" == "" ]; then
                                                            handleError "Unable to determine disk start location.";
                                                    fi
                                                    adjustedfdsize=`expr $sizefd '+' $win7part2start`;
                                                    parted -s $hd -a opt u kB mkpart primary ntfs ${defaultpart2start}B ${adjustedfdsize}kB &>/dev/null;
  

•                                           parted -s $hd u kB set 1 boot on &>/dev/null
  

•                                                   parted -s $hd u kB set 1 boot on &>/dev/null
                                            else
                                                    handleError "Invalid partition count.";
                                            fi
  

And a follow up. Ran into another issue, perhaps related… For windows 7 images (osid 5) it seems to think partitions
always start at 2048sector (line 145) But my windows 7 partiton in this case starts at 63…

So after the recreate partclone.ntfs doesn’t see a ntfs filesystem on sda1 (starting at sector 2048)
If I set it back to 63 it works.

Why 2048 and not ${partitionStart}s sectors?

Or maybe osid 5 should not be in that if statement at line 137, just do the fancy resize stuff for win8/8.1 ?

Or maybe I have a goofy win7 system? (the system in question is a virtual box VM)

Any clue bat application woul be appreciated. Im very far from a windows grue, so my understanding of what is normal or abnormal in the partition set up is quite sparse…

Hi folks, me again.

fog 1.1.1, SL6.5 base, fresh install not upgrade.

Yesterday I posted about getting into debug mode to try to figure out why an upload I was trying was failing.
Thanks for the clue!!

I think I have figured out the cause of the problem, but I am not 100% about my solution.

In the init filesystem bin/fog.upload line 139ish $win7part1start is set via the output of a parted command.
In my case the returned value after the parted | sed pipeline was 32.3

So when that gets used to:

adjustedfdsize=expr $sizefd '+' $win7part1start;

expr returns: expr: non-numeric argument

and adjustedfdsize is blank…

so later when the partition is “adjusted” (IE deleted and recreated) the recreate fails so when
it trys to partclone, partclone fails to find $part ( /dev/sda1 in my case)

Im thinking a wipe of decimal place and beyond could be the fix

Line 139 changed to:

win7part1start=parted -s $hd u kB print | sed -e '/^.1/!d' -e 's/^ [0-9]*[ ]*//' -e 's/kB .*//' -e 's/\..*$//'

and line 148 too.

Thoughts?

Howdy folks,

Been a long time fog .[23]x user. Today decide to try out fog 1.1.1 with a fresh instal. SL6.5 base OS.

Installed it, and then created a image. Modified our tftp server for one host (our clonemaster7 server)
to use:

next-server x.x.x.x;               // the fog server
filename "undionly.kpxe";

and finally got that server to boot to pxe menu. (learned a (little) bit about ipxe along the way, kinda cool)

Oh, and figured out that a kernel download was needed and that if you have a 64 bit machine
( cpuid --ext 29 ) you are going to get a 64 bit kernel and (more important if you manually set to a 32 bit kernel that you know used to work fine with .32, then you will get a 64 bit initrd with that 32 bit kernel

But when I tried to upload the image, I wound up having a failure, something about /dev/sda1 not available or some such. It blasted by to quickly.

With trouble like this in .32 world I would go to “debug” boot to see what the kernel sees and sometimes
even “run the fog steps” manually till we found the “problem”. But there doesn’t seem to be a debug option anymore. I imagine with some (more) research on ipxe I could find a incantation to have it boot the kernel with init=sh or somesuch, but I thought I would ask here first if there is a “official” way to do this.

Thanks! Having high hopes for 1.0 branch, and hoping not to have to backport windows 8 crap into 0.32

E