Image task ignoring location settings for nodes and hosts
I have been reading these forums for a couple of weeks now working on a multi-node FOG implementation.
I have 2 locations, both schools that share a 100mb VPN link between sites.
The sites have different subnets, but share a Windows AD Domain.
I have a Master FOG server at Location A, and 2 storage nodes at Location B.
I have tried to set storage groups, and locations associated with those groups, but I cannot seem to get hosts at Location B to image from the nodes locally.
When I set a hosts location to Location B, I can start the image task, and it will attempt to start imaging over the WAN from Location A. However, if the Master FOG server at Location A is currently using all its slots, the host will move on to the local node at Location B.
I have upgraded to 1.5.0, had these issues, then upgraded again to 1.5.2. Finally, I rebuilt all the servers because I was having issues with the log viewer and I think I may have corrupted some settings.
CUrrently, all 3 servers are running Ubuntu Server 17.10 LTS, with FOG 1.5.2 installed. Both Storage Nodes at Location B have been installed as Storage Nodes, and not full servers.
Replication seems to work fine - I can also PXE Boot from either the storage nodes, or the Master FOG server via the VPN. Yet, using either PXE server, I still have the issue with the Hosts pulling images from the Master FOG server.
Any suggestions or recommendations would be great. I have watched the video in the wiki, and read sever posts from Tom and some VERY detailed posts from Wayne - it seems I have set things up correctly, but I must be missing something.
Thanks in advance!
I wanted to update you, I have performed fresh installs of Ubuntu on all 3 servers, and installed the latest working branch version of FOG, 126.96.36.199
This time I did not import my hosts and old images and started fresh by re-registering hosts to test with. This time I am not having issues with the location plug in. Perhaps it was the old host data which was causing these issues.
Anyways, I appreciate all your help and time, and your effort on this wonderful product which makes many other peoples lives much easier.
Thank you for your help Tom. I am going to re-build with a fresh OS install and install the working branch directly tomorrow morning.
After downgrading to 1.4.4, I was still having issues. I looked in the SQL database, found some entries in the location and locationAssist tables which were incorrect. I cleared these entries after uninstalling the plugin. I reinstalled it and still had the issue with the wrong node being selected.
I will let you know if I continue to have issues after a fresh rebuild on all nodes.
I’ve done some testing in an attempt to see if this problem is still occurring, and i’m not seeing anything wrong since my last push to the working branch.
If you’d be so kind as to repull and install, let me know if you’re seeing the same or still having an issue.
rhyspowell last edited by rhyspowell
So, I decided to try and roll back to 1.4.4 since I see a lot of posts saying the location plugin worked fine in that version. I rolled everything back after purging PHP and Apache packages, and I am still having the same issue.
When I initiate an image task for a Host in location B, it will pull from the master node in Location A, if I initialize a second task, it will use the correct node however.
I also removed and re-installed the location plug in thinking that might help too, and deleted and re-created the Locations as well. Still no luck.
Is there anything else I can do to troubleshoot or check?
Also, I am not sure if it is related to the boot, because when the task is created in the Web GUI, the “Running Tasts” page shows which node the Host is working with and demonstrates the issue
I’ve found a problem with how the storage node was being found during boot menu. I don’t know if this translates, still, to the location plugin or not, but hopefully it’s addressed. If not, please let me know and I can take a look at it more directly.
rhyspowell last edited by rhyspowell
Just as an update, I tried adjusting the storage groups and assigning them to the locations, but I’m still having the problem with nodes pulling from the wrong location.\
In fact, after trying some different scenarios, they appear to be ignoring the storage group associations as well. Seems that the hosts will just pull from any node that has the image regardless of the location or storage group assignment.
OK, I upgraded the nodes and master server to the working branch just now.
188.8.131.52 is the version they are all running now.
Still experiencing the same issue, I will adjust my storage groups and nodes to try the virtual servers with 0 host slots now that that issue is resolved.
Thanks for your help!
Would you mind checking out the working branch to see if this is fixed. I don’t know what’s causing the problem, but I imagine some of my other code was doing something. I found a couple issues that I did make corrections for. I just don’t have a reliable means to test different nodes and the location plugin together right now.