How to avoid locking a crashed machine on the DDB screen

A brief word of explanation: DDB is the kernel debugger for Open BSD. Whenever your Open BSD machine/server crashed it drops into DDB by default.

But what if your machine is a server, and you don't want it to be locked for 24 hours before someone can move into the data center and reboot it? Don't laugh this happened to me!

Well, then, your only solution is to make sure the kernel configuration is set properly. Here is the default:

# sysctl -a | grep -i panic                                                                                                                             

If the ddb.panic kernel parameter is set to 1, the Open BSD kernel will drop into DDB when it encounters a panic.

To set it to 0 and disable the DDB screen, do this:

# sysctl ddb.panic=0
ddb.panic: 1 -> 0

Check again:

# sysctl -a | grep -i panic 

Make the change permanent by editing the /etc/sysctl.conf file and uncomment the following line:

# grep -iv ^# /etc/sysctl.conf                                                                                                                          
ddb.panic=0                     # 0=Do not drop into ddb on a kernel panic

Please note that when ddb.panic is set to 0, the kernel is supposed to save a crash file that contains more information on what occured to raise a panic. I haven't tested this yet, but it promises to be interesting.

See Also: