Posts made by ch3i

@Raymond-Bell Hi, is it new hosts or hosts that are disappeared that you want to register ? I’ve a same problem with my FOG server, some mac address are missing in the mac address table, I’ve to delete all hosts which are not in the mac address table to register it.

@baboun Hi, Have you multiple network interfaces ?

Try to search on forums, there are lot of topic about that error : https://cse.google.com/cse/publicurl?cx=013831856522586480408:mlbymhymgxs&q=Starting to restore

@Tom-Elliott said:

@ch3i much appreciated but I meant the code for the sed statement not the function code that needs to be altered.

https://forums.fogproject.org/topic/5952/erase-of-exports-file/11

@pixomondo said:

Hello,

I have moved my ESXi guest image to a new host with an SSD. Since then my image starts to fail to upload with a read DMA error at 95%. I’ve setup a new server with the latest version of FOG but I still get the same error. Has the image been corrupted during transfer? Is the SSD at fault?

Here is a screenshot of the error (this is from version 0.32):
http://imgur.com/hHStgPX

Thanks,
Felix

Hi, can you try to upload/download a new image to ensure that is or not a problem during transfer ?

@Wayne-Workman said:

@Tom-Elliott has added a new option in /opt/fog/.fogsettings

blexports=

1 means to rebuild /etc/exports
0 means do not rebuild it.

https://forums.fogproject.org/topic/5975/dhcpd-conf-file-gets-rebuilt-on-every-install

It’ll be better to have it automatically in the installer like the @olivierl solution.

@Tom-Elliott In lib/common/functions.sh

configureNFS() {
    echo -e "$storageLocation *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)\n$storageLocation/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)" > "$nfsconfig";
    dots "Setting up and starting RPCBind";
    if [ "$systemctl" == "yes" ]; then
        systemctl enable rpcbind.service >/dev/null 2>&1 && \
        systemctl restart rpcbind.service >/dev/null 2>&1 && \
        systemctl status rpcbind.service >/dev/null 2>&1
    elif [ "$osid" -eq 2 ]; then
        true
    else
        chkconfig rpcbind on >/dev/null 2>&1 && \
        $initdpath/rpcbind restart >/dev/null 2>&1 && \
        $initdpath/rpcbind status >/dev/null 2>&1
    fi
    errorStat $?
    dots "Setting up and starting NFS Server..."
    for nfsItem in $nfsservice; do
        if [ "$systemctl" == "yes" ]; then
            systemctl enable $nfsItem >/dev/null 2>&1 && \
            systemctl restart $nfsItem >/dev/null 2>&1 && \
            systemctl status $nfsItem >/dev/null 2>&1
        else
            if [ "$osid" == 2 ]; then
                sysv-rc-conf $nfsItem on >/dev/null 2>&1 && \
                $initdpath/nfs-kernel-server stop >/dev/null 2>&1 && \
                $initdpath/nfs-kernel-server start >/dev/null 2>&1
            else
                chkconfig $nfsItem on >/dev/null 2>&1 && \
                $initdpath/$nfsItem restart >/dev/null 2>&1 && \
                $initdpath/$nfsItem status >/dev/null 2>&1
            fi
        fi
        if [ "$?" -eq 0 ]; then
            break
        fi
    done
    errorStat $?
}

@Wayne-Workman said:

@ch3i How would I use those functions from CLI?

I suppose I can make a bash script during debug and call those functions to run, then run ntfsfix on all the partitions in the same script.

You can modify the fog.upload to call these scripts before uploading.

@Uncle-Frank said:

As far as I can see in the fog.upload script those two functions are only used with image type set to resizable.

I think we have to enable it on each upload resizable or not.

@Wayne-Workman Hi, There is a function in func.sh (in the init) to remove pagefile and hiberfile :

# $1 is the device name of the windows system partition
removePageFile() {
    local part="$1";
    local fstype="";
    if [ "$part" != "" ]; then
        fstype=`fsTypeSetting $part`
    fi
    if [ "$fstype" != "ntfs" ]; then
        echo " * No ntfs file system on ($part) to remove page file"
        debugPause;
    elif [[ "$osid" == +([1-2]|[5-7]|9|50) ]]; then
        if [ "$ignorepg" == "1" ]; then
            dots "Mounting partition ($part)";
            mkdir /ntfs &>/dev/null;
            ntfs-3g -o force,rw $part /ntfs;
            if [ "$?" == "0" ]; then
                echo "Done";
                debugPause;
                dots "Removing page file";
                if [ -f "/ntfs/pagefile.sys" ]; then
                    rm -f "/ntfs/pagefile.sys" >/dev/null 2>&1;
                    echo "Done";
                else
                    echo "No pagefile found";
                fi
                debugPause;
                dots "Removing hibernate file";
                if [ -f "/ntfs/hiberfil.sys" ]; then
                    rm -f "/ntfs/hiberfil.sys" >/dev/null 2>&1;
                    echo "Done";
                else
                    echo "No hibernate found";
                fi
                resetFlag "$part";
                umount /ntfs;
            else
                echo "Failed";
            fi
            debugPause;
        fi
    fi
}

And ntfsfix too :

# $1 is the part
resetFlag() {
    if [ -n "$1" ]; then
        fstype=`blkid -po udev $1 | awk -F= /FS_TYPE=/'{print $2}'`;
        if [ "$fstype" == "ntfs" ]; then
            dots "Clearing ntfs flag";
            ntfsfix -b -d $1 &>/dev/null;
            echo "Done";
            debugPause;
        fi
    fi
}

Not sure if the two functions are used during upload process, @Developers ?

@Wayne-Workman said:

This is with Fast Startup and Hibernation turned on.

I’ll try that : ntfs-3g -o remove_hiberfile /dev/sdaX /tmp/mount_point to remove the hibernate file.

@kverkiss Hi, which version of fog ?

@Wayne-Workman said:

@ch3i Not yet, thank you for reminding me. I’ll try to do it today. I actually need to.

Hi @Wayne-Workman, I’ve tested it on a Windows 8.1 with “Fast Startup”, it’s functional. Could-you test it to confirm my result ? If it’s OK, it would be interesting to add it to FOG upload process.

@Asthea said:

Ups, sry, wasn’t specific enough I guess, I meant GUI widgets.

hi, is it to have some widget on the dashboard ?

@Uncle-Frank said:

Yes, the installer script creates a new empty (overwrites) exports-file. It’s not easy to write fail-proof script code that checks to see if other entries (or old FOG entries) are in that file. So we better create a clean new one. If you have a good idea how to better do this you are more than welcome to give us a hint.

Please make a backup copy of your file before running the installer. @Tom-Elliott Well, maybe we should do that like we do with /tftpboot…

Maybe use comments before/after the fog exports, like that :

/custom/config1

#FOG settings
/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)
#END FOG settings

/custom/config2

On each update store all other lines not in these comments

@Tom-Elliott said:

I don’t understand the problem.

He have custom sharings in his “exports” file which are overwrite after each FOG update, I think lol

@robin-masters said:

@ch3i

Thx…for this.

Who maintained this page?https://wiki.fogproject.org/wiki/index.php/Using_FOG_with_an_unmodifiable_DHCP_server/_Using_FOG_with_no_DHCP_serverundefined

Under the text: "Make a symlink for the undionly.kpxe file so dnsmasq can find it. " eventually it`s better to use here the relative link it should work for all systems?

@Wayne-Workman , it’s for you.

It’s work for all system, I think it’s a tftpd server limitation.

@robin-masters said:

@ch3i

Hey lol it works… tried before again with ln -s /tftpboot/undionly.kpxe /tftpboot/undionly.0 and it definitively dont work.

ln -s undionly.kpxe undionly.0 works now… oO

Tried pxe boot from Dell Optiplex 790 and 7010

@robin-masters said:

@Tom-Elliott Yes it`s correct i use dnsmasq only with a proxy setup. Like i said a symlink to undionly.0 seems dont work … but a copy… seems illogical, eventually i have misspelled but i copy the line from the manual
ln -s /tftpboot/undionly.kpxe /tftpboot/undionly.0
After that a PXE boot was not possible it cant find the file. With cp /tftpboot/undionly.kpxe /tftpboot/undionly.0 it works…

Hi, try relative link instead absolute.

ln -s undionly.kpxe undionly.0

@Wayne-Workman Hi, did you have time to test ?

@the_duke In your /images/dev/ have you something ? it’s was there is a “R” here >> ls -alR /images