0

I've been placed in charge of an OS X 10.5 server (which is trapped at 10.5 because of its PPC architecture.)

Unfortunately (perhaps around once a week - frequently enough to be really, really annoying) the DHCP and VPN servers loose all of their settings (reverting to factory defaults), and are switched off by the system. I was tired of being alerted to the problem by irate users without internet or VPN access, so I wrote a small cron script using the serveradmin utility to restore the settings should serveradmin status dhcp return "OFF". This is a real bandaid solution (still results in some downtime) and I'd like to fix the actual problem, but I can't find anyone with similar issues. I want to move the DHCP function away from this server eventually but for various reasons this isn't an option at the moment.

Here's what the DHCP log (from /var/log/system.log) yields:

Sep 30 08:21:28 server bootpd[2710]: can't open /etc/bootptab
Sep 30 08:21:28 server bootpd[2710]: server name server.ourcompanydomain.com
Sep 30 08:21:28 server bootpd[2710]: interface en0: ip 192.168.0.110 mask 255.255.255.0
Sep 30 08:21:28 server bootpd[2710]: subnets: Failed to convert 'router': Invalid IP address

I'm not sure whether this is relevant or not.

Any assistance would be greatly appreciated!

Christopher
  • 103
  • 2

1 Answers1

0

I've concluded that this is most likely to occur when someone leaves a user logged into the console and leaves the Server Admin application running and connected to the server. I've seen the behavior in a variety of locations, but it didn't hit me until I did it to my own server at my office.

Leave Server Admin open and connected on the console = DHCP and VPN die, and their settings reset to defaults.

Chris Dawe
  • 16