Wake on LAN over different VLANS
-
Good afternoon,
I have trying to test wake on lan through fog over different vlans. On the same vlan as the fog server, wake on lan works perfectly, but when I try wake on lan over a different vlan thats when the packet seems to not go through. For example:
My fog server IP is 10.1.0.119, if I try to wake on lan a pc with an IP address of 10.1.x.x, wake on lan works perfectly. I then tested computers with an IP of 10.2.x.x and an IP of 10.3.x.x, and the computer never wakes up. I used the same computer in all locations so the settings are the same just something is blocking the packet from going through.
Anyone have any ideas?
-
What version of fog? The trunk version has a plugin designed specifically for this. It’s called wol broadcast.
-
@szecca1 And ensure you have configure your switchs for WOL across VLAN.
Cisco example : http://www.cisco.com/c/en/us/support/docs/switches/catalyst-3750-series-switches/91672-catl3-wol-vlans.html -
I apologize for the long delay, I have been very sick. I am running FOG version 1.2.0 and am able to Wake on LAn with a program call Goverlan.
I saw this on the wiki site
https://wiki.fogproject.org/wiki/index.php/How_to_use_the_wakeonlan_perl_script_instead_of_etherwake.exe
wasnt sure if that was relevant.
We are running Avaya and Nortel switches through our district and have opened the ports, at least we believe unless we are missing something. The wake on lan works on the same vlan but crossing vlans is wake it stops. -
@Tom-Elliott The wol broadcast plugin is not in the plugins for some reason so I cant install that
-
-
@ch3i Yes, thats what is confusing. I tested today turning off a computer in a different vlan and then the computer woke right up with my wake on lan from Goverlan but wont do the same in FOG
-
@szecca1 Goverlan is on the same VLAN than FOG Server ? I ask you for that because in my network only one VLAN can WOL on other vlans.
-
@ch3i Yes I am running goverlan from my machine which is on the 10.1 VLAN and FOG is a VM that also running on the 10.1 VLAN
-
@szecca1 when I refer to Trunk, i am referring to “SVN”, the current/latest development side of fog. Trunk is just a shorter term than constantly saying here’s 3559, or 3560, or so on and so forth. If you can Upgrade to trunk (the link is embedded here to the wiki article) you will have the WOLBroadcast Plugin. I did make a custom version of it for my work place as we had similar problems, but I don’t feel comfortable spreading that out with 1.2.0 having so many bugs in its nature. Hopefully you understand.
-
I suspect you are running some sort of traffic management device and the reason your desktop can wake on lan across vlans is that it is authenticated to the traffic shaper and the linux server is not.
-
@Tom-Elliott I completely understand and fully appreciate all your help. I just updated the trunk and got the wolbroadcast plugin installed. Is there anything else I need to do or technically now it should work?
@Joseph-Hales The Fog server is probably more authenticated than my computer is as it no longer needs to authenticate to out iPrism and my computer still does every so often. -
@szecca1 You need to instlal and activate the plugin of course. The last step is add in all your broadcast addresses with names (for simplicity all around)
-
@Tom-Elliott Ok so I installed and activated the plugin and when I click on it i am getting this:
You’re saying all I have to do now is name the broadcast whatever i want and then give the broadcast IP of what VLAN it is?
-
@szecca1 yes
-
@Tom-Elliott Ok awesome I will give that a try tomorrow and let you guys know how that works. I appreciate all the help.
-
That’s a pretty awesome plugin… for real…
I’m going to use that when I setup FOG in vmware at our central site.
-
@Tom-Elliott I apologize in advance but where would I get this broadcast ip? Would this be the gateway ip from each building? I’m not sure where to pull this broadcast IP from?
-
You basically jot down any valid IP for a network, then jot down your subnet mask.
Convert both to binary, with each of the four octets separated by a decimal as normal.
Then, for each ‘0’ in the subnet mask, you’d flip that to a 1, but you’d still use the network portion of the valid IP, those binary bits would just “fall through”
So… example time…
Subnet mask: 255.255.240.0
IP 10.2.3.8
Binary IP Subnet:
00001010.00000010.00000011.00001000
11111111.11111111.11110000.00000000(Client IP network portion: 00001010.00000010.0000 Client portion: 0011.00001000)
the fall through numbers for the network address:
00001010.00000010.0000000.00000000 (Where 1 is set in the subnet mask, binary from the client ‘fall through’ to form a network address)
This gives us a network address of 10.2.0.0The broadcast address is the network address combined with all-on client bits.
00001010.00000010.00001111.11111111
(binary in the subnet mask set to ‘0’ ‘fall through’ to the broadcast address as ‘1’)Convert to decimal: 10.2.15.255
So, 10.2.15.255 would be the broadcast address for that IP and that subnet mask.
And, that’s correct, that’s my particular building’s broadcast address.
-
@Wayne-Workman The 10.2.3.8 address that you used in your example, where did you get that from? Is that just a random IP of a client in the 10.2 VLAN or is that something in particular?
