I made a clone of my existing production OOD server (it’s a vmware vm) to test v3 and get ready for the migration in 3 weeks. I changed the hostname and updated ood_portal.yml then ran the update_ood_portal script. I’m able to hit the new server but it immediately generates a dex login URL that points to the original production server. I checked the ondemand-dex config.yml and it’s pointing to the right hostname. I cleared the Apache cache in /var/cache/httpd and now it bypasses ondemand-dex and gives me a 500 Internal Server error.
Is there an easy way to do this? Or should I just install from scratch?
Hey sorry for the trouble. Which OS are you using?
Are you following the upgrade instructions in the docs:
If so, did you run the
ondemand-dex update as well already?
I’m running Rocky Linux 8.8. I haven’t run any updates yet, was trying to get it to work first. /etc/ood/config/ood_portal.yml is the only place I need to change the hostname, right?
Yeah, that would be where the
hostname would be changed. I ask about the
ondemand-dex because the docs call out that if you are using
dex you will also need to update
ondemand-dex and I wondered if this was what is causing your auth error.
Okay I fixed it. I had to go get a new certificate and also fix /etc/httpd/ssl.conf which was pointing to my production server.
I’ll try an upgrade path tomorrow.