Nodes not behaving properly on a rebuild
Backstory: We have a master set up on a VM that is running really slow on the web interface. I’ve tried to troubleshoot it a large part of the year. We also had a problem with the nodes and the size of the VM for them. The thought was that we’d build a new master on a dedicated machine with a different IP and move the nodes over during the summer. The new master is up. We’re starting to rebuild the nodes. It’s not working quite right.
In the web interface, I can see the nodes under Storage Management but not under the Dashboard nor under FOG Configuration, Kernel versions. We are also unable to capture an image. The first client error was ftfp relate. I have that working now (mismatched tftp and fogstorage passwords and permissions on /images). Now we’re getting the PXE-T01: File not found error as the client is trying to boot. We handle our own DHCP and we have it set up the same as we have the other master set up. I’ve spent a good part of a day looking over settings, some of that time with our networking person. I don’t think we need to set up the dnsmasq since we can configure our DHCP options.
Any advice would be appreciated. Thanks in advance.
@wayne-workman Thank you. So far we seem to be working through them but we aren’t quite back to normal yet.
Wayne Workman last edited by
@lpetelik Good deal. If you hit other issues, feel free to post here so we can try to help.
@wayne-workman Thanks. We do have the nodes showing now however, I think the info you gave will be helpful for the nodes pointed to the previous master that are still functional.
Our problem was in DHCP in the location where the master was at. I was scouring the nodes and missed changing it at the location where the main server was grabbing it’s DHCP Scope Options.
Wayne Workman last edited by
but not under the Dashboard nor under FOG Configuration, Kernel versions.
Any errors in the apache log? On CentOS 7 that should be
/var/log/http/error.logAlso about what @Quazz said, the nodes would need to be updated with the new master’s IP address - that should be a matter of updating
/opt/fog/.fogsettingsand re-running the installer on them. Also - the Node FOG Version and the master FOG Version must match or things won’t work right.
Thank you for the advice.
Yes, the node sites have DHCP on the win servers pointing to the fog nodes.
I did see that document after I started the migration a slightly more complicated way. If I can’t make this work right, I’ll redo it all and go with this document.
Did you change the DHCP options to point to the new IP?
Go over https://wiki.fogproject.org/wiki/index.php?title=Migrate_FOG if you haven’t already. (don’t forget to also update /opt/fog/.fogsettings)
As for the nodes, I’m guessing that while the main server has got the correct information, the nodes themselves aren’t informed about the IP of the new server.
I forgot to give version info.
The master is on CentOS7, current build with FOG 1.5.3 ( I do see there is another update).
Nodes are on Windows Server 2012 r2 in Hyper-V.