HP9000 won't boot

[Maarten van Maanen]

We have a HP9000 D370 with 10.20 and a PCI btlan4 J3515A card which was not recognized by the system. We installed some patches with no succes. Then we manually installed a patch from the support CD. Now the system won't boot anymore. When I get it is single user mode it will panic after a while. When trying a reboot I get:

/sbin/rc[61]: syntax error at line 3 : ')'unexpected.
INIT: Command is respawning too rapidly
Will try again in 5 minutes
Check for possible errors
id: cons "/usr/sbin/getty console console
and then:
id: ems4 "/etc/opt/resmon/lbin/p_client

When I start in single-user maintenance mode I cannot mount any volumes. Mount -a give me a "No such device or address" on all volumes.
With lvlnboot -v I get "Volume group not activated"

What can I do to correct this problem ?

[John Palmer]

You need to activate your volume groups before you can mount your filesystems.

vgchange -a y

The init message is caused by a faulty script in /sbin/init.d

[Andreas Voss]

Hi,

in maintenance mode the vg00 is not activated.
Use vgchange -a y vg00
But be carefull: no not change to other init state! reboot instead

Regards

[John Palmer]

Further to my previous post. Line 61 in /sbin/rc is where it executes /etc/rc.config. This executes every script in /etc/rc.config.d so one of the files in there is incorrect. Can you check what file(s) have been modified recently?

[Kofi ARTHIABAH]

Yes.. I had a similar problem once - and it was because my oracle people had set up an rc script in the /etc/rc.config.d directory! the /etc/rc.config.d directory should contain scripts that only initialise variables that are then used by the scripts in /sbin/init.d/

Boot to single user mode, then go to the /etc/rc.config.d directory and look for any scripts that are large (usually, your scripts will be larger than the others) the files in this directory should only contain comments and lines of the sort:

START_MYAPP=1

if you have any scripts anything else ... it might be suspect - move it to a different directory and restart.

[Maarten van Maanen]

Tried what you've set with vgachange -y a and run fsck. Had to do that several times though before some volumes could be mounted. Checked the rc.config.d dir, moved the new hpbase100 file to somewhere else and renamed the .old one.
This systems will boot now, but the btland4 card is still not working.
One other question I still have.
/ is mounted under /dev/root instead of /dev/vg00/lvol3. Is this usual under 10.20?

[John Palmer]

Maarten,

This is a known problem after maintenance mode. Just recreate /etc/mnttab by removing it and then recreating it with mount.

rm /etc/mnttab
mount

Regards,
John

[Maarten van Maanen]

Mount -p also gives me /dev/root on / instead of /dev/vg00/lvol3

[Cheryl Griffin]

Might this be the problem? You posted that you have a "HP9000 D370 10.20 PCI btlan4 J3515A".

D class would have an HSC bus, not PCI.
We often see calls where installing the wrong hardware, software, and/or patches can prevent the system from booting.

The D class has should have an HSC (FF EISA) or EISA bus not a PCI. The product number would be J3516A (vs. J3515A which should be for a C or J class machine with an HSC bus. The C and J class machines have 1 RJ45 and 1 AUI port and use the btlan4 driver.)

The D class HSC/FF EISA should have 2 RJ45 and 2 SC/MMF ports. It uses the btlan4 driver. The HSC/EISA would have only 1 RJ45 port and use btlan0 driver.

For each bus and lan card, there are associated patches. For a D class 10.20 machine (this list includes dependencies):
PHCO_10947 libHcurses cumulative patch
PHNE_16298 traceroute patch
PHCO_16591 fsck_vxfs(1M) cumulative patch
PHKL_16751 SIG_IGN/SIGCLD,LVM,JFS,PCI/SCSI cumulative patch
PHKL_16957 Physical dump devices configuration patch
PHKL_17858 Fix for mount/access of disc sections
PHNE_18061 arp general patch
PHCO_18563 LVM commands cumulative patch
PHNE_19566 ping(1M) cumulative patch
PHKL_19886 GSCtoPCI PCI to PCI bridge cumulative patch
PHSS_19916 OV ITA 3.20 Server Patch
PHKL_20611 Correct process hangs on ufs inodes
PHNE_20834 cumulative ARPA Transport patch
PHNE_20893 HSC 100BT lan cumulative patch
PHCO_21186 Cumulative SAM/ObAM Patch
PHNE_21216 LAN products cumulative Patch
PHKL_21595 VxFS (JFS) mount, fsck cumulative patch
PHKL_21661 lo_realvfs panic fix, Cumulative LOFS patch

[Maarten van Maanen]

The ioscan output is given below:
lan 1 10/12/1/0 btlan4 UNCLAIMED UNKNOWN PCI(10110009) -- Bu
ilt-in #1

Concerning the list of patches you gave. Isn't there a cumulative patch for 10.20 that will include all of these ?

[P V Patel]

Hi,
You may activate volume group with vgchange -a y command. e.g.
vgchange -a y /dev/vg00

After activate all required volume group
run fsck command, then
mount /usr and /var filesystems. You may get logical volume with following command:
more /etc/fstab
Now you can execute sam or you can edit file.
Identified recently modified files in /etc/rc.config.d directory with ls -lt command and correct modified scripts.
After correcting problem (with script) try to reboot or try to boot in multiuser mode with init2 command.

[Tom Danzig]

FYI, EISA cards on D3XX are not supported if you're running 64 bit OS.

[P V Patel]

Hi,
init2 will be init 2 in previous reply.
Punjabhai V Patel

[Cheryl Griffin]

This is a 10.20 system, so 11.00 EISA issues are not the problem here.

The patches are for different things: SAM, ARPA, LAN, etc. All are required due to dependencies, so there is not 1 megapatch available.