posted to this is not a weblog on wednesday, march 7th, 2007
tagged nerd, openbsd, and never commented on

after installing a openbsd 4.1 on the netra last night and taking it to a datacenter today, it suddenly decided to stop booting from the disk and went back to its annoying net boot loop.

for whatever reason, these machines won't accept a break (~#) from tip/cu while in this sequence. even issuing "#." to escape into lom and telling it to issue a break won't work. there doesn't appear to be any point in the boot sequence where issuing a break works to interrupt the boot before it starts looking at the network.

[...]
Boot device: net  File and args:
Using Onboard Transceiver - Timeout waiting for AutoNegotiation Status to be updated.
Timeout reading Link status. Check cable and try again.
Link Up.
Timeout waiting for ARP/RARP packet
Timeout waiting for ARP/RARP packet
Timeout waiting for ARP/RARP packet
#.
lom>break
Timeout waiting for ARP/RARP packet
Timeout waiting for ARP/RARP packet
[...]

the solution is to break into lom, then tell it to boot into forth mode (the "ok" prompt) and then change the boot device.

#.
lom>bootmode forth
lom>reset

LOM event: +0h30m19s host reset
[...]
Probing Memory Bank #3 256 Megabytes
Aborting startup sequence because of lom bootmode "forth".
Input and output on ttya.
Type  fexit  to resume normal startup sequence.
Type  help  for more information
ok setenv boot-device disk
boot-device =         disk
ok nvstore
ok reset