Master not delegating imaging to storage nodes and no tasks showing up on Dashboard?

[danieln]

Hello,

Before I begin, here are my specifications:
OS: Linux/Debian
FOG Version 1.5.8

I’m kind of at a loss here. I have a FOG setup with a Master and eight storage nodes. As of this morning I noticed that whenever I started imaging, all of the computers would connect and deploy from one of the storage nodes. Thinking this was odd, I disabled this node in the Storage Node settings of the FOG Dashboard just to see what would happen. Then the computers all began connecting to only the master node and began to oversaturate it. I currently have ~25 computers connected to FOG and imaging from the master node and they’re all pulling about 150MB/min (very slow compared to how it usually runs).

What even more bewildering is that there are no tasks showing up on the FOG dashboard either. It’s still showing that theres 68 free spots available and no activity at all. I also only have an 11 Max Client limit on the master and like I said, I easily have 25 computers connected to it pulling from the master. I have tried rebooting the master and every node and no luck.

All Linux and FOG services seem to be up and running after running sudo systemctl

Any ideas on what may be going on?

Thanks in advance,

[Sebastian Roth]

@danieln said in Master not delegating imaging to storage nodes and no tasks showing up on Dashboard?:

What even more bewildering is that there are no tasks showing up on the FOG dashboard either.

What do you mean by that? You don’t see any tasks in the task list page? Did you schedule tasks?

[Sebastian Roth]

@danieln Please see if you can figure out what’s going on in the database:

shell> mysql -u fogmaster -p
...
mysql> use fog;
...
mysql> SELECT * FROM tasks WHERE taskStateID <= 3;
...
mysql> quit

[danieln]

@Sebastian-Roth Thank you for the reply!

I can manually schedule tasks and they show up in the task list. But when I pxe-boot computers to the FOG network and depoy an image, it doesn’t show any active tasks for the deployment even though it’s deploying 25 copies of the same image from the master server.

[Sebastian Roth]

@danieln Do I get this right? You’re saying it just doesn’t show up in the dashboard graphs?

[danieln]

@Sebastian-Roth It does not show up on the dashboard or in the task list, no.

But the bigger issue is that all of the computer imaging is only pulling from the Master when I have 8 other storage nodes. It is flooding the master when the master is only supposed to have an 11 client maximum. And the deployment task seems to not be created although it is deploying.

[Sebastian Roth]

@danieln This sounds really strange. As well because one time you say “[tasks] show up in the task list” and then “does not show up on the dashboard or in the task list, no”. So what?!

Please run the database query and post output here.

Do you use plugins like “location”, “capone” or any other plugin?

[Sebastian Roth]

@danieln Possibly some colleague “forced” the tasks to run while all your other storage nodes are seen as offline? That could cause what you are describing.

[danieln]

@Sebastian-Roth Thank you again for your reply.

I believe I found the fix. The colleague was not using “quick registration” to register the MAC addresses into the database before imaging. It seems that the Master needs to have the MAC address of each client in the database to delegate imaging tasks to each node.

I have a full bench of computers now (registered) and they are now pulling from all 8 nodes equally and tasks are showing. Thank you again.

[Sebastian Roth]

@danieln Ohhh yes, didn’t even think about this. Sure if you deploy straight without registering then it’s a totally different story.