Strange behaviour, v240 stalls

Hi,

Ive got two V240 that acts strange from time to time.

Yesterday evening they stalled, I connected to the consol and nothing happened. No OK-prompt or nothing, when pressing Enter to try to get some response the marker jumped down a line but nothing more happened.

I then rebooted it and everything worked, then when I disconnected the consolcable the v240 went to OK-prompt, very annoying. I noticed that when disconnecting the session (reflection) and pulling the consolecable the v240 went to OK-prompt. but if I pulled the consolecable without disconnecting the session the machine was working fine.

Anyone have any ideas?

T&R,

Paul

[700 byte] By [pf3016] at [2007-11-25 22:49:07]
# 1
When the problems occur are you able to get the the SC prompt using the '#.' key sequence. If so then the output from the showlogs command might show something.
stumoor at 2007-7-5 17:04:35 > top of Java-index,Sun Hardware,Servers - General Discussion...
# 2

The rebooting behaviour is due to the way you have the /etc/default/kbd file configured. This is setup this way so if the system hangs you can just disconnect the console / keyboard (dependant on the system) and it will drop to the ok prompt. This way you could type sync and generate a coredump if needed.

If you don't want this behaviour though you can edit it as per the instructions in the man page kbd

andthings at 2007-7-5 17:04:35 > top of Java-index,Sun Hardware,Servers - General Discussion...
# 3

So that others might understand a bit better ...

The 'system hang' symptoms were completely unrelated to how the console reacted to their cable connection.

The drop-to-the-OBP behaviour is indeed influenced by settings in the /etc/default/kbd file of their Solaris environment.

It's a Solaris <i>feature</i>, though not always perceived as a benefit.

Paul noted in his initial post that they got to the OK prompt after they tried to check the system and was therefore it was not a cause of the hang symptoms.

If they had used telnet to check the system, it would have no influence on how the console reacted.

Paul would need to investigate the system completely, using Stuart's suggestions, and then decide if they need to induce a system core.

Bill at 2007-7-5 17:04:35 > top of Java-index,Sun Hardware,Servers - General Discussion...