Due to some changes at our company, we lost the password for the root and sms-svc account on the primary system controller (E25K).
We connected a serial console to the system controller but we never managed to break the session to bring it to the OK prompt.
(tried: ~#, [ENTER]~[CTRL] b, ~[CTRL][BREAK], [CTRL][BREAK], [CTRL] a b, [SHIFT]-#, sending break from terminal option, unplug/plug keyboard from serial console client, etc.)
As the system controller failover is not enabled, we can't rely on the other system controller... I know it's bad practice and it's really dumb to loose the passwords... but probably you all know how it goes in these huge IT companies where systems are managed by teams of people.
My questions are:
1.) Is there any other way to break into the OBP?
2.) Is there any way to force the secondary SC to take over the role of the primary, while not having access to the primary?
Thanks for your input sparcguy. But do you mean the PS2 connector of a client pc/laptop that has a serial connection to the system controller? Because there is no PS2/USB connector on a E25K system.
Ah .. E25k I'm not familiar, we don't have that kind resources here to maintain such an animal in our 40+ server zoo, pity tho.
But yes Solarius,
The older servers use PS2 connector keyboards at the console, when you remove and replace the connector it will send the server crashing straight down to ok prompt.
I know that the newer servers all use USB keyboards, but I've not tried this method with USB keyboard yet
Thanks for the answers guys. I contacted Sun regarding this, this is what they answered:
>If you do not shutdown the domains, then you will not be able to reset the passwords
>of the root and sms-svc accounts on the main SC.
>
>As you have no failover enable and so the spare SC cannot takeover the function
>of the main SC.
>
>You cannot simply break into the main SC as that would cause all the domains
>to crash.
This is bad news... as we want to avoid any service interruption by all costs as it hosts services for more than a dozen clients. But I guess we'll have no choice.
Strange though that a monster like the E25K still has to depend on a reboot; that Sun has no way to force a failover.
you CAN force a failover but you need the passwords to do so... losing passwords is always a bad thing. so i think it's not the fault of sun... isn't it?
That's just because the slackers who should have checked the config of this server when it was handed over to production didn't do their job properly. Two words: ITIL bureaucracy