FOG 1.5.6 Multicast problem?!

[kratkale]

Hi,
till now I’ve used FOG 1.5.4 with multicast. I’ve installed a new debian based machine and the FOG Server is now running on a HP Elite 8300 (I5 8GB RAM 120GB Samsung SSD and 1.2TB HDD mounted for the images). The FOG version is the really brand new 1.5.6.
Installation was working perfectly. Upload of a new image as well.

Multicast deployment starts perfectly with a group of 25 PCs. But then all the PCs are stuck with the partclone screen but there is no data transfer (the red bars of the block process do not appear).

Now I clone 90 PCs just with normal Deployment which takes several hours compared to 8 Minutes with multicast FOG 1.5.4 …

I’ve checked Kernel 4.19.36 (booting), 4.19.6 (booting), 4.19.1 (booting) and 4.1.2 (kernel panic). That didn’t help.

The client computers are all:
System Manufacturer LENOVO
System Product 10MNS1QP00
System Version ThinkCentre M910t

Thank you for your advice

Alexander

[george1421]

This is the first I’m hearing about issues with multicasting with FOG 1.5.6.

This issue needs to be sorted out, but I’m getting the impression from your post that you have an urgent need. What you need to get moving forward again is the init(s) init.xz and init_32.xz (FOS linux vritual hard drive) from FOG 1.5.4. You can keep the kernel at the latest.

I have not checked but I suspect you can get the older inits from the tarball: https://github.com/FOGProject/fogproject/archive/1.5.4.tar.gz

they go into /var/www/html/fog/service/ipxe directory. Just rename the 1.5.6 inits to save them and move the 1.5.4 inits into place.

[Sebastian Roth]

@kratkale Before releasing 1.5.6 I have tested multicasting (with two clients) and it worked perfectly fine. Have you restarted the FOG server after installing 1.5.6?

[kratkale]

@Sebastian-Roth Good evening,
yes I restarted the debian server several times;
I’ve tried now a multicast with two PCs - no succsess

[kratkale]

@george1421 hi, thank you both for your support. I cannot find init.xz and init_32.xz in 1.5.4.tar.gz.

I’ve still the old ssd of the 1.5.4 in spare.

I’ll get the init(s) from ssd tomorrow and check it.

Thank you so much for your quick reply.

[Sebastian Roth]

@kratkale The old inits won’t help you I am fairly sure. But if you want to give it a try you can download those here: 1.5.5 and 1.5.4

Please take a look at the multicast log and post the full contents here. You find that log in /var/log/fog/fogmulticast.log

[kratkale]

@Sebastian-Roth said in FOG 1.5.6 Multicast problem?!:

/var/log/fog/fogmulticast.log

[05-03-19 1:30:08 pm]

=== ==== ===== ====
=== ========= == === == ===
=== ======== ==== == ==== ===
=== ======== ==== == =========
=== ==== ==== == =========
=== ======== ==== == === ===
=== ======== ==== == ==== ===
=== ========= == === == ===
=== ========== ===== ====

===== Free Opensource Ghost ======

============ Credits =============
= https://fogproject.org/Credits =

== Released under GPL Version 3 ==

[05-03-19 1:30:08 pm] Interface Ready with IP Address: 127.0.0.1
[05-03-19 1:30:08 pm] Interface Ready with IP Address: 127.0.1.1
[05-03-19 1:30:08 pm] * Starting MulticastManager Service
[05-03-19 1:30:08 pm] * Checking for new items every 10 seconds
[05-03-19 1:30:08 pm] * Starting service loop
[05-03-19 1:30:08 pm] | This is not the master node
…
The last line is repeated 6000 times with different times …

[kratkale]

@Sebastian-Roth
What I’ve done:
-) HD
took the old image hdd put it in the new system.
/dev
/postdownloadscripts
*.mntcheck
i copied from the new installation to the old disk

-) old csv files
Export in the old system:
+) users
+) Host
+) Groups
+) Snapins
+) images
+) printers
import to the new system

[Sebastian Roth]

@kratkale Seems like it does not find the right network interface. Please double check the FOG config settings in the web UI for the multicast network interface setting.

As well, which Linux OS do you use?

[kratkale]

@Sebastian-Roth
it is a brand new installed debian
root@FOG:/images# cat /proc/version
Linux version 4.9.0-9-amd64 (debian-kernel@lists.debian.org) (gcc version 6.3.0 20170516 (Debian 6.3.0-18+deb9u1) ) #1 SMP Debian 4.9.168-1 (2019-04-12)

normal deployment of single computers works very well!

[kratkale]

@Sebastian-Roth
http://192.168.0.196/fog/management/index.php?node=about&sub=settings

[Sebastian Roth]

@kratkale I tested on Debian as well, strange. Will take a look at the code when I get home in half an hour.

[Sebastian Roth]

@kratkale Ahhhh, just remembered that we had a similar issue, see here https://github.com/FOGProject/fogproject/issues/268

We ought to find a better solution to this…

[kratkale]

@Sebastian-Roth
This aproach is cousing a boot loop " Unable to locate image store …"

/var/www/html/fog/lib/fog/fogbase.class.php, jump to line 2226 and comment the first three lines of the getIPAddress() function. Make it look like this:

    /**
     * Get IP Addresses of the server.
     *
     * @return array
     */
    protected static function getIPAddress()
    {
//        if (count(self::$ips) > 0) {
//            return self::$ips;
//        }
        $output = array();
        exec(
            "/sbin/ip -4 addr | awk -F'[ /]+' '/global/ {print $3}'",
            $IPs,
            $retVal
        );

[Sebastian Roth]

@kratkale Let’s try to take a new approach to this. Please edit /lib/systemd/system/FOGMulticastManager.service, goto line 18 that looks like this:

After=syslog.target network.target mysql.service

Change it to look like this:

After=syslog.target network-online.target mysql.service

Then reboot your FOG server and check /var/log/fog/fogmulticast.log again.

As well you might run systemd-analyze plot > systemd-deps.svg and upload that SVG graphics file to a file share and post a link here. On my Debian test VM it looks like this:

The FOG* services start late on that test system (latest Debian in virtualbox VM) anyway and that is without the above mentioned change. Would be interesting to see if that looks different on your hardware server. Possibly the network interface card on your hardware just takes a little longer…

[kratkale]

Now the addrees is here:

…
[05-04-19 10:25:00 am] Interface Ready with IP Address: 127.0.0.1
[05-04-19 10:25:00 am] Interface Ready with IP Address: 127.0.1.1
[05-04-19 10:25:00 am] Interface Ready with IP Address: 192.168.0.196
[05-04-19 10:25:00 am] * Starting MulticastManager Service
[05-04-19 10:25:00 am] * Checking for new items every 10 seconds
[05-04-19 10:25:00 am] * Starting service loop
[05-04-19 10:25:00 am] * No tasks found!
…

[Sebastian Roth]

@kratkale Then schedule another multicast task and see if it’s working. I guess we are on the right way with this.

Earlier on you said:

This aproach is cousing a boot loop " Unable to locate image store …"

Where did you see this message?

[kratkale]

@Sebastian-Roth
just the messeage you see in the pic. I’ve still this problem!

There are no images anymore???
I ve to check the mounting …

[kratkale]

@Sebastian-Roth
ok - I lost the image I’ve done new on this configuration.

The old images (importet with csv file) are still available in Web GUI

in /images I can find all the folder: old an new …

There was a mistake in the fstab file - so the hdd was not mounted

[kratkale]

ok I took now an old image
it is working but:
-) Erasing current MBR/GPT Tables takes a looooooooooooooooong time
-) Cloning works much faster 12,7GB/min instead of 8GB/min
no mistakes with the old image