Posts made by Phenomenon

I am having a little difficulty understanding how the image replication works. I was trying to setup FOG to store all the images in the /images/ folder and only have the images that were tagged with a specific storage group to move to a storage node.

Currently we have a Master server with its storage group set to default. We also have another master setup on the web ui of the server assigned to a storage group for a school. We also have a storage node setup with the storage group set as the school. The server was successful in replicating images but instead of only pulling the images that were tagged with the specific storage group for the school it pulled the entire /images/ folder.

Does anyone have any advice as to what we can try to resolve this issue? If there is anymore information needed just let me know and I will do my best to reply quickly.

To give you an idea of our layout we currently have our FOG server running on a VM with 2 physical nodes on a different VLAN within our building. We are using Ubuntu and FOG 1.0.0. The server had images on it before we connected the nodes, and once they were connected we were successfully able to replicate images from the master to the nodes. We were also able to get the graphical display for storage info on the web UI for both nodes. We are also able to deploy images from the nodes successfully.

The issue we are having is when we attempt to capture an image from a computer on the same VLAN as the nodes we get the FOGFTP rename error. We have found that when it is being captured the data is being stored on a node instead of the master, and when it goes to rename the file it looks in the /images/dev/ on the master and can obviously not find it since the data is being stored on a node first. We can move the folder from the node to the master manually while it is looping through trying to rename/move the folder and it will work but that is not a solution to the problem but a work around. From our understanding the way FOG works, it is supposed to capture to the master server first then replicate to the nodes.

Does anyone have any advice as to how we can set the image to be captured to the FOG server first by default instead of the node?

[IMG]http://imageshack.com/a/img842/8017/j0v9.png[/IMG]
[IMG]http://imageshack.com/a/img841/7333/ueef.png[/IMG]
[IMG]http://imageshack.com/a/img834/5533/2i3m.png[/IMG]
[IMG]http://imageshack.com/a/img844/124/wecb.png[/IMG]

What we are attempting to do is have a centralized distribution at our main office and have it replicate out the the servers at our schools. We have had success in replicating images to our locations but the issue we are seeing is the master is replicating images from a different storage group.

We currently have something similar to:
Master -> IP for Central server -> Storage group (CO) - > Storage location /images/
Master -> IP for Central server -> Storage group (<School>) -> Storage location /images/
Storage Node -> IP for School server -> Storage group (<School>) -> Storage location /images/<School>

[IMG]http://imageshack.com/a/img838/3552/fpkf.png[/IMG]

We assigned an image to the <School> storage group and captured the image. Instead of the Master server replicating just the images for the storage group they were assigned it replicated every image we had in the /images/ folder.

Is this working as intended and I am simply misunderstanding the point of assigning a storage group to a server and image?

What is the point of assigning a Master/Node and Image to a storage group? Is there not a way for the replication to distinguish which storage group the images are assigned?

We have moved the nodes to the same VLAN as the Master and changed the storage location IP config to the Masters IP address. This seemed to resolve the issue for a little while. After the weekend we came back and were running a few more tests before we implemented this solution for our servers. When we tried to capture again it started uploading to the storage node again instead of the master. We are hoping to get this up and running for our summer imaging throughout our district. We will be imaging 4000-6000 new computers and roughly 10000 existing computers.

If anyone has any advice as to what we could test it would be greatly appreciated.

bump