Fail to boot up after halting the machine

[Angus]

My HP server is running in HP-UX 10.20. When I run the command of "shutdown -r", it restart smoothly and everything goes fine. However after I halt the machine (shutdown -h) and want to restart it by powering it on, it fails at a point of executing /sbin/bcheck.sh (seems checking the volume group). I don't know why. The machine seems can only be booted up after interacting the ISL and boot-up with singer user mode first follow by the command of "shutdown -r". Do anybody got any idea to fix it?

Thanks.

[Rita C Workman]

This is a 'tuffie'...but here are my thoughts (for whatevery they are worth).
I saw on another post someone with an L class server complaining of similar...no resolution reached there it seemed. You didn't mention your server class.
Now, since you get up with a -r, I tend to rule out software or volume group issues. If the system had a problem reading the volume groups, it would have it on a reboot too... But here's some info on what bcheck is--for your notes if you want it:
http://www.csee.usf.edu/cgi-bin/man-cgi?bcheck+1

So I am leaning toward hardware. Since you power off then the hardware must be re-recognized..and it's on the -h your seeing the issue. First thing I would do on my next attempt is double check my keyboard and monitor cables..I might even swap out with another keyboard,monitor that I know has no issues. It never ceases to amaze me how a simple keyboard can create such headaches. Then I'd ensure that the cables and everything connecting to disks is right. In other words I would (at least till I was comfortable) confirm peripherals and connections to peripherals.
Now this may not be the answer....but it is something to at least rule out...

Regards,
/rcw

[John Bolene]

A few more diagnostice woukld help.

How about doing a dmesg and see if it reports anything strange, and look in syslog also.

It may be that you have a bad super block.

We have several machines that will not restart communications on a shutdown -r. The FDDI card does not get reintialized. We have to do a halt and then cycle power. Pain in the butt for remote admins.

It could be that one of your cards is having a problem initializing after a power down. Or it could be a block of memory has problems.

[melvyn burnard]

I would hav ethought there might be some error or warning message that gets displayed on the console. If so posting it woul dhelp.

[A. Clay Stephenson]

Hi Angus,
I basically agree with Rita that it's hardware. That's the essential difference between a warm and cold boot. One thing that might reveal unrecogized hardware is to do an ioscan while you are up normally. Then boot to
single user and do an ioscan. (ioscan is in /sbin so no bheck stuff needs to be done.) I also suspect that this problem will only grow worse with time. One thing that could cause problems like this is a slighty out of spec power supply that does just fine except for startup load like spinning up disks.

My 2 cents worth, Clay

[paul courry]

Marginal disks may also take a long time to come up when power has been interrupted. As a rule I don't power down disks unless a) it is absolutely necessary and b) I have a backup or c) I feel lucky