could not start in runlevel 3 and always stops in runlevel 1

[Ricky_4]

Hi All,

HPUX 11.00, which is installed in L class, could not start normally. It is always stopped in runlevel 1. If we are trying to switch to runlevel 3 using init 3, it only hangs. How could we isolate it why it could not start in runlevel 3?

Thanks in advance.

Adriatico

[Patrick Wallek · Accepted solution]

Check the /etc/rc.log file and to see which startup script it is hanging on then work from there.

[Ricky_4]

Hi Patrick,

We observed that partitions on /etc/fstab is not mounted in runlevel 1 and could not execute some of the commands. Please assist us to overcome this so we could review /etc/rc.log.

Thanks.

Adriatico

[Denver Osborn]

what exactly isn't being mounted? is it one of the root vg filesystems? /usr?

Is there possibly a bad/failed disk?

What's the output of...

mount -a
vgdisplay /dev/vg00


-denver

[Ricky_4]

We tried executing "mount -a" in single user but only few filesystems from fstab are mounted.

How could verify if there are failed/bad disk? but thru light indicator of disk all disks are normal.

Thanks.

[Denver Osborn]

We really need to know more about the system to figure out what filesystems aren't being mounted and why.

What's not being mounted. Is this box using LVM or non LVM layout? please post what from the /etc/fstab was not mounted.

Assuming it could be a problem with the root VG, also post the output of "vgdisplay -v /dev/vg00".

-denver

[Ricky_4]

All OS and applications partitions are using LVM layout.

And dont have any ignite ux for this server.

var, tmp and applications filesystems were not able to mount.

[Denver Osborn]

ok... so the next question is... where is /var? is it part of root vg or was it moved to another vg (same for /tmp)?

What VG do /var, /tmp and your app filesystem belong to? Look at the /etc/fstab entry... /dev/vg??? What are the VG and LVol names.

If they're not in vg00, try to activate the VG they belong to.

If the problem was simply that another VG that contained your app, /var and /tmp wasn't activated and all is well after a vgchange -a y /dev/mystery_vg ; mount -a ; init 3... then you should be looking at the /etc/lvmrc file. If it isn't an MC/SG node, then check the /etc/lvmrc and make sure it is set to AUTO_VG_ACTIVATE=1

I'm going to bed soon, so I hope this last bit of info helps!

-denver

[Patrick Wallek]

Well, in run-level 1, or single-user mode, what you are seeing is normal. The only filesystem that gets mounted is /stand. And if you do a 'mount -a' you will only see the filesystems associated with VG00 be mounted. That is because the other VGs have not been activated yet.

What happens if you try to bring the system up in stages?

If it stops at single-user mode, then do an 'init 2' to go to run-level 2 and see what happens. Watch the output on the console (or web console or lan console) and note any errors. Watch for the normal startup scripts to run.

If that is successful, then do an 'init 3' and see if that works. If it stops somewhere, make note of the script that is running when it stops. You can try doing a | (That is hold down the CTRL key and SHIFT key and do a pipe | symbol) to interrupt the problem script.

If you can get the system up, you can then check the /etc/rc.log and /var/adm/syslog/syslog.log for errors.

[Ricky_4]

Hi All,

Thank you all for your help.

We have figured out our problem.

Thanks again.

Adriatico

[Patrick Wallek]

So what WAS the problem? It is helpful to post the solution so other people can potentially benefit from it.