Latest FOG 0.33b

jbsclm

I can see what is going wrong in the capone code, I’ve got the management plugin working correctly, and am looking at the service code, but I’m missing something obvious at the moment.
I’ll post the amended versions when I sort it out.

Troye Johnson

Tom it is working I am still getting errors in the log file on the client side but the snapins are deploying.

Tom Elliott

access_log is not an error. It expected so you know what people are accessing. It’s how I figured out snapin’s weren’t working.

It’s also how I was able to determine you had multiple NIC’s on that system.

Tom Elliott

On the client, there is the one I think is more a problem with how fog determines the directory structure. It tries to find the alo-bg.png file from c:\program files\fog\images\alo-bg.png, but on 64 bit systems the real location is:
c:\program files (x86)\fog\images\alo-bg.png.

I’m still working on the other errors, though I think they’re ok, for now. Still trying to figure out how Usercleanup is supposed to work.

Tom Elliott

r1134 released. Fixes many service things. All seems to report properly now and supports the hosts that have multiple nics. As long as the macs are registered with the host in the fog gui, all should work properly. (Meaning wireless stuff should work as well.)

Tom Elliott

r1136 released.

Partclone upload tasks (multi-part) uploads the correct (hopefully) for the partition.
ext{2,3,4} = partclone.extfs -c
ntfs = partclone.extfs -c
fat{16,32} = partclone.vfat -c
anything else (for now) uses
hfp,btrfs,xfs=partclone.dd
(It’s important to note that partclone.dd does not require -c option in upload, which is why I’ve labeled them as you see above.)

Should help with the Win 8 differences in partitions.

Deploy tasks are done with:
partclone.restore rather than partclone.${fstype} so we know it restores using the correct type, rather than having me guess which are which.

blkid updated in the init.gz to allow displaying of the partition types/uuid/etc…

Minor changes, again, to fog.css to allow display of the kill icon’s. I must have made an accidental edit to the file, but it broke something. It’s fixed now.

I have not, yet, implemented gpt stuff, but will get on it. I don’t have anything to test with, so I’ll be guessing, using of course other people’s suggestions.

Tom Elliott

r1137 released.

Had to make tweak to blkid in fog script. Should work more fluidly with this method.

From the commit comments: “Make (hopefully) final tweak in partclone stuffs on fog script. Not done updating fog script, just hopefully done with partclone tweaks.”

I hope you know that just means I’m done, at least for now, with partclone methods. I will continue editing the fog script, and I hope to use partclone for dd (raw) imaging types in the future.

Troye Johnson

Hey Tom,

For the multiple snapin deployment could you make a way to select multiple files to deploy and not have to deploy them all also what would you need to test gpt.

Tom Elliott

I wouldn’t be able to have a method of deploying multiple files. Though I’d imagine (if we had a parsing method in windows) you could create a package (zip,msi,…) and use args to unzip/install/etc…

I don’t know what I’d need, I don’t have anything with gpt. There’s already a couple people looking into it as they are actually in need of it, which is useful. I hope they don’t mind helping me out in this regard.

Troye Johnson

I was thinking I can build you a vmware server giving you full access if that would help any?

joacliba

With the latest revision, I am able to backup a multipartition disk, but when I try to restore, it only says “task completed” but nothing is done.

Tom Elliott

What kind of multipart? Single or all?

joacliba

Single, one disk multiple partitions.

Tom Elliott

Is this is multi-boot system (Windows and Linux?)

joacliba

Yes. Well, NTFS and ext partitions, as windows isn’t installed yet.

Tom Elliott

That’s a kink I’ve yet to figure out. It has to create the partitions before it tries to restore the partitions. I don’t know how this is completed as the scripts used for the deploy process are:
[code] elif [ “$imgType” = “mps” ]; then
# Restore multipartion for single drive
echo -n " * Looking for Hard Disks…“;
#disk=fogpartinfo --list-devices 2>/dev/null | cut -d' ' -f 1
getHardDisk;
echo “Done”;
echo " * Using Hard Disk: $hd”;

                if [ -n "$hd" ]
                then 
                    # check image
                    if [ -d "$imagePath" ]
                    then 
                        tmpMBR="${imagePath}/d1.mbr";
                        if [ -f "$tmpMBR" ]
                        then 
                            echo -n " * Restoring MBR...............................";
                            dd if=$tmpMBR of=$hd &>/dev/null
                            if [ "$osid" == "50" ]
                            then 
                                fdisk $hd  &>/dev/null << EOFLINUXFIX

w
EOFLINUXFIX
fi
echo “Done”;

                            echo -n " * Checking hard disks.........................";
                            runPartprobe;
                            echo "Done";
                            parts=`fogpartinfo --list-parts $hd 2>/dev/null`

                            diskLength=`expr length $hd`;
                            for part in $parts
                            do   
                                partNum=${part:$diskLength};
                                echo " * Processing Partition: $part (${partNum})";
                                sleep 1;
                                imgpart="${imagePath}/d1p${partNum}.img";
                                if [ "$osid" == "50" ]; then 
                                    echo " * Preparing linux filesystem...";
                                    parted -s $hd -a opt &>/dev/null << EOFPTD

mkfs
yes
$partNum
fat32
q
EOFPTD
runPartprobe;
sleep 10;
fi
if [ -f “$imgpart” ]; then
if [ “$mc” = “yes” ]; then
if [ “$partNum” == “2” ] && [ “$osid” == “6” ]; then
udp-receiver --nokbd --portbase ${port} --mcast-rdv-address ${storageip} 2>/dev/null | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
elif [ “$partNum” == “3” ] && [ “$osid” == “6” ]; then
udp-receiver --nokbd --portbase ${port} --mcast-rdv-address ${storageip} 2>/dev/null | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
else
udp-receiver --nokbd --portbase ${port} --mcast-rdv-address ${storageip} 2>/dev/null | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
fi
else
export FOGSTATS=“1”;
if [ “$partNum” == “2” ] && [ “$osid” == “6” ]; then
pigz -d -c < $imgpart | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
elif [ “$partNum” == “3” ] && [ “$osid” == “6” ]; then
pigz -d -c < $imgpart | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
else
pigz -d -c < $imgpart | partclone.restore -O $part -N -f 1 2>/tmp/status.fog
fi
fi
else
echo " * Partition File missing: $imgpart";
sleep 9;
fi
sleep 1;
clearScreen;
echo " * Image Restored";
done
if [ “$osid” == “50” ]
then
echo " * Restore Linux swap paritions:“;
swaps=LANG=C fdisk -l | grep swap | cut -d' ' -f1 2>/dev/null;
IFS=’
';
for x in $swaps
do
echo " * Restoring partition: $x”;
mkswap $x &>/dev/null
done

                            fi
                            echo "";

                            echo " * Task complete!";
                            echo "";

                        else
                            handleError "Image Store Corrupt: Unable to locate MBR.";
                        fi
                    else
                        handleError "Unable to locate image store.";
                    fi
                else
                    handleError "Fatal Error: Disk device not found!";
                fi[/code]

You’ll notice it doesn’t create the partitions itself, it uses the MBR of the system to generate the partition table for this block.

My guess is something is wrong with the DD command creating the backup MBR during the upload process.

joacliba

I don’t understant the osid variable, if it’s multipartition and it can hold multiple os, for what is $osid?

I can’t test until monday, but, it’s supposed to exist a [FONT=Consolas]/tmp/status.fog where I can check errors?[/FONT]

Tom Elliott

/tmp/status.fog is the location for the progress information on the client side. fog.statusreporter looks at the data in the file, sends it to the fogserver then clears it out for the next iteration.

The osid, I think, is more for specialized cases. If its a pure linux system it needs to create the parts in the right places before upload can begin, hence the OSID in place for 50. Otherwise, it doesn’t appear to care.

fabritrento

[quote=“Tom Elliott, post: 21867, member: 7271”]r1137 released.

Had to make tweak to blkid in fog script. Should work more fluidly with this method.

From the commit comments: “Make (hopefully) final tweak in partclone stuffs on fog script. Not done updating fog script, just hopefully done with partclone tweaks.”

I hope you know that just means I’m done, at least for now, with partclone methods. I will continue editing the fog script, and I hope to use partclone for dd (raw) imaging types in the future.[/quote]

i tryed to do a upload of a pc with:
sda1 ntfs
sda2 ext4
sda3 swap

the first partition works, and also the swap.

the ext4 partition don’t work, i think the cause is because blkid command returns ext2 ext3 ext4 and the partclone.extfs command is not called.

So i created 3 symlink in /usr/sbin on the client:

ln -s partclone.extfs partclone.ext2
ln -s partclone.extfs partclone.ext3
ln -s partclone.extfs partclone.ext4

but not works.

is possible to modify the script to make 3 different log in /var/log/partclone_[DATETIME].log ? to debug? i find only the 3th launch of partimage in the log

Tom Elliott

can you give me exact output of blkid for ext4 stuff?

Latest FOG 0.33b

172

11.6k

17.1k

154.6k