03-13-2023 09:58 AM
All of our switches have a primary and a secondary radius server defined by its DNS name.
Example
configure radius mgmt-access primary server hq-nac-01.domain.com
configure radius mgmt-access secondary server dr-nac-01.domain.com
We moved the dr-nac-01.domain.com to another datacenter and so it had to change its IP.
I can ping dr-nac-01.domain.com from a few test switches and they reply back with the proper new IP address. However when I test by disabling the NAC service on our HQ server... the switches just hang and timeout and never log on. I ran wireshark on the DR radius server and I don't even see the switch trying to reach out.
So in the switch I do a show radius and it has the proper host name but the wrong IP address. But yet if I ping the name like I said above, the proper IP is returned.
I reran the command: configure radius mgmt-access secondary server dr-nac-01.domain.com 1645 client-ip ipaddresshere vr VR-Default but it stills hows the OLD IP in there.
The DNS servers have been updated. How can I correct this on about 15 exos switches?
03-14-2023 05:26 AM
While quite intrusive, you could try to restart the "netTools" process to see if this helps... "restart process nettools"
06-15-2023 06:05 AM - edited 06-15-2023 06:05 AM
I never got around to trying this, I ended up deleting and reconfiguring radius on all 15 switches. It seems like it only does a DNS lookup of the radius server the moment you configure it. It doesn't abide by DNS TTL's and recheck every so often.