FOG 1.2.0 multicast issues

[TRSD]

Since it looks like /images is read-only in the exports file should I edit that file for rw or is there a different way of changing it? Or is it not necessary?

[Tom Elliott]

is that from fog server or the box where /images is moutned from?

[TRSD]

That is from the box where the images are stored.

[Tom Elliott]

Okay,

I’d say copy the exports from there to your FOG Server, if they don’t match already.

In that exports only set:
[code]*(rw,sync,no_wdelay,no_root_squash,insecure)[/code]

[Tom Elliott]

Then restart the NFS services on both the hosting images box and the fog server.

[code]sudo service nfs-kernel-server restart[/code]

[TRSD]

They both match. Which should I change, fog server or image box?

[Tom Elliott]

The image box.

[TRSD]

done. restarted nfs kernel servers and FOGMulticastManager and still not working

[Tom Elliott]

Then I’m at a loss as to what this issue is.

One last thing to try if at all possible:
[code]umount /images
touch /images/.mntcheck
touch /images/dev/.mntcheck
scp -r 10.1.0.56:/images/PeteTest /images/
chown -R root:root /images
chmod -R 777 /images[/code]

If you have enough room of course. Then try redoing the tasking, does all work?

[TRSD]

touch: cannot touch `/images/dev/.mntcheck’: No such file or directory

[Tom Elliott]

mkdir /images/dev
touch /images/dev/.mntcheck

[TRSD]

This says it will take about 18 minutes. I will leave it running but I have to leave the office now. I really appreciate your help with this and will report back in the morning with a status.

Thank you very much.

[Tom Elliott]

okay, thank you,

[TRSD]

Unfortunately, we did run out of space on this so we can’t test on the one server. Do you think permissions would be an issue since unicast imaging does work?

[Tom Elliott]

Yes.

Unicast access the file system on /images as root user.

This means that it’s the same permissions as if it were on the nfs.

The Multicast tasks, initially, get started from the www-data user. So if permissions are read only, it can’t execute anything on them.

[TRSD]

Would it make sense to change the owner of PeteTest to www-data for testing purposes?

[Tom Elliott]

No,

Because the /images directory is still mounted read only.

[TRSD]

Also, why wouldn’t I change the /etc/exports folder on the fog server to match the image box:?

/images *(rw,sync,no_wdelay,no_root_squash,insecure)
/images/dev *(rw,sync,no_wdelay,no_root_squash,insecure)

[TRSD]

exports file, I mean.

[TRSD]

I have done a fresh install of Ubuntu 14.04 LTS and SVN and restored database.
Now the multicast process doesn’t stop but it still doesn’t multicast.
Here is the log:

[07-24-14 1:55:48 pm] * No tasks found!
[07-24-14 1:55:58 pm] | Task (1) Multicast Group Quick Deploy is new!
[07-24-14 1:55:58 pm] | Task (1) /images/PeteTest image file found.
[07-24-14 1:55:58 pm] | Task (1) 2 client(s) found.
[07-24-14 1:55:58 pm] | Task (1) Multicast Group Quick Deploy sending on base port: 58570
[07-24-14 1:55:58 pm] CMD: cat /images/PeteTest/rec.img.|/usr/local/sbin/udp-sender --min-receivers 2 --portbase 58570 --interface eth0 --full-duplex --ttl 32 --nokbd;cat /images/PeteTest/sys.img.|/usr/local/sbin/udp-sender --min-receivers 2 --portbase 58570 --interface eth0 --full-duplex --ttl 32 --nokbd;
[07-24-14 1:55:58 pm] | Task (1) Multicast Group Quick Deploy has started.
[07-24-14 1:56:08 pm] | Task (1) Multicast Group Quick Deploy is already running PID 2520
[07-24-14 1:56:18 pm] | Task (1) Multicast Group Quick Deploy is already running PID 2520

I have checked all of the options we have previously gone through in this thread and omping does work for multicast. Any ideas since we can get this far?