Multicast not starting

[gerald.stachl]

I use Fog Trunk 7593 (but i had the same problem also with older versions):
I created a Multicast Testgroup with just 1 Notebook. When i try to start a multicast session, the notebook boots into imaging, but never starts multicast.
When i look into the mysql tables - it seems, that the client registers at the fog server, but nothing happens. Any idea what i can test next?

Thanks for any advice,
Gerald

image url)

[Quazz]

What do you mean boots into imaging but never starts multicast?

Where exactly does it go wrong? Before or after partclone? Any pictures you could take?

[Wayne Workman]

He probably means it hangs at the partclone screen? This is a common problem.

Here are some starting points:
https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_Downloading_-_Multicast
https://wiki.fogproject.org/wiki/index.php?title=Multicast

[gerald.stachl]

Yes it hangs at the partclone screen. I try to make some screenshots tomorrow.

I already did a lot of the tests from https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_Downloading_-_Multicast and
https://wiki.fogproject.org/wiki/index.php?title=Multicast

I’m going to check it again tomorrow

Gerald

[gerald.stachl]

@Quazz image url)

There are not entries in multicast log, except them created during my tests with udp-sender and udp-reciever on the server and client console.

I also tried to create a Multicast Image and boot the client to join this multicast session. Same result.

i think that the client connects to the server (msClients=1 (was 0 before), but msSessClients ist still 0 - is that correct?

[Quazz]

@gerald.stachl First thing to try is turn off the firewall on the FOG server, next up would be to check your network infrastructure with the data on the wiki. (I don’t know a ton about that, personally)

[Sebastian Roth]

@gerald-stachl Please run ps ax | grep udp on your FOG server when you have the partclone screen waiting. Post the full result you get here!

[gerald.stachl]

@Sebastian-Roth
root@imaging:~# ps ax | grep udp
45546 pts/0 S+ 0:00 grep udp

Seems udp-sender ist not running.

My Fog Server ist virtualized (HyperV)

[Tom Elliott]

@gerald.stachl is the interface correct for this node? If you run ifconfig on the fog server and look at the GUI under Storagr Management->[The same node you ran ifconfig on]->Interface, do the named interfaces match?

[gerald.stachl]

@Tom-Elliott yes eth0 for both

[Tom Elliott]

@gerald.stachl Is FOGMulticastManager running?

[gerald.stachl]

@Tom-Elliott yes
Service FOGMulticastManager status is Ok

[Tom Elliott]

@gerald.stachl Can you post the /opt/fog/log/multicast.log file?

[Tom Elliott]

I’ve added some code to more dynamically get the master interface so we don’t continue to have these issues (hopefully).

Basically, what my latest commit does, FOG will no longer need the Storage Node interface being set properly for our multicast tasks to operate. It will attempt to figure out which interface to use based on the host/ip of the storage node master. With any luck, this will mean no more of the lost udp sessions, though I can’t be 100% sure it won’t still cause issues. I’m relatively certain this change should help us with at least not losing the process that performs the tasking.

[gerald.stachl]

@Tom-Elliott
multicast.log:
12:02:27.250206 Using mcast address 234.2.0.26
12:02:27.250265 UDP sender for /opt/fog/.fogsettings at 10.2.0.26 on eth0
12:02:27.250284 Broadcasting control to 224.0.0.1
12:03:27.018433 Starting transfer: 00000009
12:03:27.039993 Doubling slice size to 162
12:03:27.043873 Doubling slice size to 202
12:03:27.062477 Transfer complete.
12:03:27.062513 Disconnecting #0 (10.2.2.88)
12:03:27.062522 Disconnecting #1 (10.2.2.152)
12:03:27.062531

the entries are from my tests with upd-sender on the console. There are not antries from mutlicast sesssions from fog.

I did the update via git - still same situation.

Where are the main config files now? /opt/fog/services/etc/config.php has only a definition of WEBROOT

[Wayne Workman]

@gerald.stachl Just a basic question here, you know that when you do a git pull to update your fog installation files, you then need to re-run the installer correct? You can run it like this: ./installfog.sh -y

[gerald.stachl]

@Wayne-Workman
Yes. After “git pull” i did
./bin/installfog.sh
passwd fog (i use a different PW for the fog user)

[Sebastian Roth]

Do you actually see the FOG services running on the server? ps ax | grep FOG?

[gerald.stachl]

@Sebastian-Roth
root@imaging:~#
ps ax|grep FOG
49456 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGImageReplicator/FOGImageReplicator
49467 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGSnapinReplicator/FOGSnapinReplicator
49478 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGTaskScheduler/FOGTaskScheduler
49489 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGPingHosts/FOGPingHosts
49760 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGMulticastManager/FOGMulticastManager
52203 ? Z 0:00 [FOGMulticastMan] <defunct>
52267 pts/0 S+ 0:00 grep FOG

i killed ps 52203 and then i rebooted the server. After that i got:

2271 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGImageReplicator/FOGImageReplicator
2324 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGMulticastManager/FOGMulticastManager
2382 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGPingHosts/FOGPingHosts
2411 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGTaskScheduler/FOGTaskScheduler
2441 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGSnapinReplicator/FOGSnapinReplicator
4428 ? Z 0:00 [FOGMulticastMan] <defunct>
4429 ? Z 0:00 [FOGImageReplica] <defunct>
4430 ? Z 0:00 [FOGPingHosts] <defunct>
4431 ? Z 0:00 [FOGTaskSchedule] <defunct>
4432 ? Z 0:00 [FOGSnapinReplic] <defunct>
4444 pts/0 S+ 0:00 grep FOG

sometimes the zombie processes disappear:
2271 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGImageReplicator/FOGImageReplicator
2324 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGMulticastManager/FOGMulticastManager
2382 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGPingHosts/FOGPingHosts
2411 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGTaskScheduler/FOGTaskScheduler
2441 ? S 0:00 /usr/bin/php -q /opt/fog/service/FOGSnapinReplicator/FOGSnapinReplicator
4868 pts/0 S+ 0:00 grep FOG

then they appear again.

On this server, unicast imaging works fine. I had some problems with the FOG Client - i had to change back to the legacy client, because the domain join didn’t work with the new client (but it worked some weeks ago - but this might be a different problem)

[Tom Elliott]

@gerald.stachl is the OS this is installed using systemd or sysvinit? If systemctl can you get output of systemctl -l status FOGMulticastManager? If sysvinit can you get any information from the Apache error logs?