Multicast problem - per host not group

ictadmin

Hi there,
When creating a Multicast task for a group of (30) hosts it looks as though the multicast task is created individually per host and when the suite of computers is booted, the download starts but works as if we had just set a unicast for all. In other words, unlike our old 0.32 setup, the server is not waiting for all hosts to get to the same ‘checkpoint’ before downloading.
Hope that makes sense.
Thanks for any help.

Tom Elliott

What version of FOG?

ictadmin

Hi sorry, we’re running fog version 1.2.0 on Ubuntu 12.04.

It appears that fog is creating an individual multicast session per host rather than one multicast session for the whole group.

If you need any further info please let us know!

Tom Elliott

When you’re scheduling the task are you telling it to do it download task or were you telling it to be a multicast task

ictadmin

We have a group set up for the computers we’re trying to multicast, we’re clicking the Multicast icon next to the group name in order to set up the multicast. So we are telling it to be a multicast task.

Tom Elliott

Are you saying you’re seeing the tasks in the active task list as individual or you saying you can start one system we’ve all the rest within the group turned off and that one system will image

ictadmin

Hopefully the attached images will help explain what’s going on a bit better.

[ATTACH=full]1214[/ATTACH][ATTACH=full]1215[/ATTACH][ATTACH=full]1216[/ATTACH][ATTACH=full]1217[/ATTACH]

When the machines are turned on from this point they will just image (they won’t wait for the others in their group) so if we were to turn one machine on it will just image.

[url=“/_imported_xf_attachments/1/1214_Multicast1.PNG?:”]Multicast1.PNG[/url][url=“/_imported_xf_attachments/1/1215_Multicast2.PNG?:”]Multicast2.PNG[/url][url=“/_imported_xf_attachments/1/1216_Multicast3.PNG?:”]Multicast3.PNG[/url][url=“/_imported_xf_attachments/1/1217_Multicast4.PNG?:”]Multicast4.PNG[/url]

Tom Elliott

Understood.

I think this is related to the “delay” issue you’re seeing on logging into the GUI.

I think what’s happening is the “delay” is causing the tasks to be generated very slowly.

So by the time one host is “tasked” the udp-sender command actually starts up. Then the next task get’s created.

ictadmin

Thanks for the suggestions, the fog sever is connected to the internet via a proxy, but the proxy port the server connects to is unauthenticated (locked down only but network acls)

I’ve just tried removing the proxy information and this seems to have fixed the issues (both slow task creation and the multicast grouping issue)

Thanks for your help!

ictadmin

One last thing on this issue, when the client boots into it’s multicast runtime we get an error saying:

/bin/fog.download: line 70: [: too many arguments

Is this normal? Just going to start googling it now!

Junkhacker

you’re not likely to find anything from google on that.
thank you for reporting that bug, and let us know if the imaging task complete successfully

ictadmin

Thanks Junkhacker, when the clients are set to multicast individually, the job completes just fine, when they are part of a group job they don’t even begin, they just sit at the blue screen as if they are waiting for another computer to join the group, have left them for an hour in this state and nothing I’m afraid!

Enrico

subscribe

Tom Elliott

[quote=“Enrico, post: 34288, member: 25297”]subscribe[/quote]

What does “subscribe” mean?

Enrico

I have the same issue:
when the clients are set to multicast individually, the job completes just fine, when they are part of a group job they don’t even begin, they just sit at the blue screen as if they are waiting for another computer to join the group, have left them for an hour in this state and nothing I’m afraid
and so I subscribe (up?) the thread to get notifications… Wrong?

Tom Elliott

So you’re not able to create a “group” multicast tasking? Or the multicast tasking never starts?

Have you verified that the FOGMulticastManager service is actually running?

ictadmin

Hi Tom,

For us the multicast task never starts, it’s as if the computers are waiting for another computer to join - or for something to kick the session off.

How can I check if the FOGMulitcastManager is running please?

Thanks,

IA

Enrico

I create the task, then I boot the 2 group members and the task starts (I can see the udpcast process and the log entry:
cat /images/test1/test1|/usr/local/sbin/udp-sender --min-receivers 2 --portbase 57050 --interface eth0 --full-duplex --ttl 32 --nokbd;

the hosts start partclone “starting to restore image…” and hang (I don’t see any traffic).
Sending a multicast task on a single host works…
Thanks for your help

Enrico

[quote=“ictadmin, post: 34293, member: 1004”]Hi Tom,

For us the multicast task never starts, it’s as if the computers are waiting for another computer to join - or for something to kick the session off.

How can I check if the FOGMulitcastManager is running please?

Thanks,

IA[/quote]
This is the log file: /opt/fog/log/multicast.log
the process is:
ps aux|grep FOGMulticastManager
root 1716 0.0 2.0 333900 20888 ? S 12:31 0:00 /usr/bin/php -q /opt/fog/service/FOGMulticastManager/FOGMulticastManager

ictadmin

Thanks for the pointers Enrico.

The mulitcast task log (multicast.log.udpcast.tasknumber) looks like this:

[CODE]Udp-sender 20120424
Using mcast address 236.16.1.48
UDP sender for (stdin) at 172.16.1.48 on bond0
Broadcasting control to 224.0.0.1
New connection from 172.16.12.2 (#0) 00000009
New connection from 172.16.12.10 (#1) 00000009
New connection from 172.16.12.17 (#2) 00000009
Starting transfer: 00000009
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
Timeout notAnswered=[0,1,2] notReady=[0,1,2] nrAns=0 nrRead=0 nrPart=3 avg=10000
[/CODE]

The multicast.log looks like:

[CODE][07-29-14 12:12:22 pm] CMD: cat /images/WIN7V13/sys.img.*|/usr/local/sbin/udp-sender --min-receivers 3 --portbase 57722 --interface bond0 --full-duplex --ttl 32 --nokbd;
[07-29-14 12:12:22 pm] | Task (45) test has started.
[07-29-14 12:12:32 pm] | Task (45) test is already running PID 22373
[07-29-14 12:12:42 pm] | Task (45) test is already running PID 22373
[07-29-14 12:12:52 pm] | Task (45) test is already running PID 22373
[07-29-14 12:13:02 pm] | Task (45) test is already running PID 22373
[07-29-14 12:13:12 pm] | Task (45) test is already running PID 22373
[/CODE]

and this goes on until we switch the machines off where the task is marked as complete.

The only other thing I can see is the /bon/fog.download line 70 error which I highlighted in a earlier post.

The FOGMulticastManagerService is running.

Thanks!

IA

Multicast problem - per host not group

166

12.0k

17.3k

155.2k