Boot into single user mode: Failed to initialize. ENTRY_INIT Status = -7

[cninham]

I am trying to boot our HP-UX 11.00 server into single user mode in order to extend the logical volumne for lvol8 (/var).

(For a text file of the "session" here below, please see the attachment)

When the server is up and running, I log in as root and from the command line, I enter:

shutdown -ry 0 (or 'reboot')

When I get the message to press any key to enter the boot menu, I press a key. At the command prompt I enter:

> BOOT PRI

then, when prompted if to Interact with IPL, I select YES.

But I get an error:

Booting...
Failed to initialize.

ENTRY_INIT
Status = -7

00000000 FFFFFFF9 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000012 00000000 00000000 00000000 00000000
00000000 454E4841 00000000 00000000 00000000 00000000 00000000 454E4841
00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000 00000200 00000000 00000000 00000000 00000000 00000000 00000000

Failed to initialize.



When I do a:

> SEA IPL

this is my response:

Path# Device Path (dec) Device Path (mnem) Device Type and Utilities
----- ----------------- ------------------ -------------------------
P0 0/0/2/0.2 intscsia.2 Random access media


Does anybody have any idea what is going on? How can I boot into single user mode...???

Thank you.

Cameron

[Scot Bean]

ENTRY_INIT Status = -7

this means that the boot device does not exist, the address specified exists but is not installed or not responding.

Is your hard drive functional?

[cninham]

I tried to boot using:

> BOOT ALT

That "worked", and I could do a

IPL> hpux -is

That allowed me to login in single user mode. However, now when I try to extend the volume group, I get an error:

# lvextend -L 2560 /dev/vg00/lvol8
lvextend: Warning: couldn't query physical volume "/dev/dsk/c1t2d0":
The specified path does not correspond to physical volume attached to
this volume group
lvextend: Warning: couldn't query all of the physical volumes.
lvextend: Not enough free physical extents available.
Logical volume "/dev/vg00/lvol8" could not be extended.
Failure possibly caused by strict allocation policy
#


Any idea why this is?

Thank you,

Cameron

[Helen French]

The error which you got normally comes when the system cannot find a boot program in the specified device (or a corrupted one). Check your primary boot path on the boot console and see if that's pointing to the right device (your primary boot disk). Also try 'search' command and see if all your disks are identified, especially the boot disk. Make sure all connections, power, cables and termination are proper.

[Helen French]

/dev/dsk/c1t2d0 - is that your boot disk? If yes, I think there is a disk failure. Check ioscan -fnC disk output and check the disk. Also use diskinfo, dd and stm tools if needed, to check the disk drive. If you find the disk is faulty, you need to replace it and then recover the data first.

[Sridhar Bhaskarla · Accepted solution]

Can you post your 'setboot' output?. There are few possibilities

1. The primary path is wrong. That's why you are getting a initialization failure.
2. Primary path may be pointing to a disk c1t2d0. But this disk is not configured as a boot disk. Probably someone might have pvcreate'd it forcefully and destroyed the LVM structure. To find it out do

diskinfo /dev/rdsk/c1t2d0
ioscan -fnC disk

and see c1t2d0 is in claimed state. DO

dd if=/dev/rdsk/c1t2d0 of=/dev/null bs=1024k

See if you get any errors. If not, then someone messed up that disk in which case 'vgcfgrestore' may help.

-Sri

[cninham]

Thank you all for your help and your reply(s),

For Sridhar Bhaskarla:

Please see here below the output of the commands you requested. I think you may be onto something. Can you help me with more details on what I need to do if the disk is in a "CLAIMED" state, please? What does this mean?

I have very little knowledge of the boot process. I would greatly appreciate it if you would provide some "step-by-step" diagnostic procedures and what to do to attempt to fix this, please.

Again, thank you ALL for your help so far.

Cameron


------------ SNIP ------------


<<< Log from tdhpws01 started January 30, 2004, 09:10:50 >>>
# setboot
Primary bootpath : 0/0/1/1.2.0
Alternate bootpath : 0/0/2/0.2.0

Autoboot is ON (enabled)
Autosearch is ON (enabled)

# diskinfo /dev/rdsk/c1t2d0
diskinfo: can't open /dev/rdsk/c1t2d0: No such device or address
# ioscan -fnC disk
Class I H/W Path Driver S/W State H/W Type Description
=====================================================================
disk 1 0/0/2/0.2.0 sdisk CLAIMED DEVICE SEAGATE ST318404LC
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0
disk 2 0/0/2/1.2.0 sdisk CLAIMED DEVICE HP DVD-ROM 305
/dev/dsk/c3t2d0 /dev/rdsk/c3t2d0
disk 3 0/2/0/0.0.0 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c4t0d0 /dev/rdsk/c4t0d0
disk 5 0/2/0/0.0.1 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c4t0d1 /dev/rdsk/c4t0d1
disk 7 0/2/0/0.0.2 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c4t0d2 /dev/rdsk/c4t0d2
disk 4 0/6/0/0.1.0 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c6t1d0 /dev/rdsk/c6t1d0
disk 6 0/6/0/0.1.1 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c6t1d1 /dev/rdsk/c6t1d1
disk 8 0/6/0/0.1.2 sdisk CLAIMED DEVICE HP C5447A
/dev/dsk/c6t1d2 /dev/rdsk/c6t1d2


# diskinfo /dev/rdsk/c2t2d0
SCSI describe of /dev/rdsk/c2t2d0:
vendor: SEAGATE
product id: ST318404LC
type: direct access
size: 17783240 Kbytes
bytes per sector: 512

------------ SNIP ------------

[Pete Randall]

Cameron,

According to that output, you're trying to boot off a non-existent device. Your primary boot path is set to 0/0/1/1.2.0, yet your ioscan output shows no such device. I would attempt booting off the alternate boot path (which appears to point to your root disk) and/or use "setboot -p 0/0/2/0.2.0 -a 0/0/2/1.2.0" to point to the correct volume and have the DVD drive be your alternate.


Pete

[cninham]

Sridhar,

It APPEARS that the boot disk is /dev/rdsk/c2t2d0, and not /dev/rdsk/c1t2d0.

Will your recommendation of:

dd if=/dev/rdsk/c2t2d0 of=/dev/null bs=1024k

fix the boot problem?

[Pete Randall]

Cameron,

No, the dd was just to check the health of the disk. In order to fix the boot problem, you'll need to change the primary boot path using the setboot command I gave you earlier.


Pete

[Pete Randall]

Cameron,

I think you used to have a /dev/rdsk/c1t2d0 disk and it has failed. What type of machine is this? How many internal drives do you have? Are they mirrored?

I'm hoping that you lost your primary mirror of the root drive. You need to replace the mirror, then rebuild the LVM configurations on it and sync the primary with your secondary.


Pete

[cninham]

Pete,

The server in question is hardware model 9000 / 800 / L2000-36, running HP-UX 11.00.

Reading the posts, it appears that the problem is a failed mirror drive -- i.e. the primary boot drive, which is [?] mirrored... Looking at the server, there are two physical 18 GB SCSI LVD drives in it. I will have to verify if the drives are configured to be mirrored...

Is there a way to verify if the actual drive has failed [PHYSICAL] or whether it is a corrupt boot image or broken mirror...?


Thanks,

Cameron

[Patrick Wallek]

I would say the chances are pretty close to 100% that you have a broken drive. The fact that you only have 1 internal hard drive (and the DVD) showing up in your ioscan output is the main clue.

Do a 'vgdisplay -v vg00' and see if your LV's defined show up as 'available/stale'. If they do, there's another clue that you've got a toasted drive.

If you haven't rebooted too many times you could also try checking /var/adm/syslog/OLDsyslog.log for any error messages relating to c1t2d0.

[Pete Randall]

Cameron,

In your other thread, your lvol8 display shows the "available/stale" indication that Patrick mentions. I would say that that guarantees that your primary drive is toast. You need to get HP to replace the drive, then rebuild the mirrors.


Pete

[cninham]

Patrick and Pete,

Thank you. Yes, the drive was "toast"... But is appears to have been a seating problem. I shutdown the server, turned off the power. Removed and reseated both drives (to their original locations), and now both drives are recognized again...

I am hoping that it the re-seating have fixed the problem, and that the drive is not suspect and likely to fail again... I will monitor the drive.

B.t.w., when I opened the cover to the L2000 server, the small green triangular LED was flashing next to the "broken" drive. If I remember correctly, it was the RIGHT side LED of the two that was flashing...

Thank you,

Cameron

[Pete Randall]

Cameron,

I find it very interesting that the drive had a seating problem. It's hard to imagine how it could have gotten unseated without someone deliberately manipulating the eject handle. I'd keep an eye on it just in case someone is experimenting.

I can't find an L-Class Owner's Guide anymore but I think this RP5450 Guide should be close enough to give you some guidance with reference to the meaning of the disk's LEDs:

http://www.docs.hp.com/hpux/onlinedocs/A5191-96018/A5191-96018.html

See the last section of the "Troubleshooting" chapter, "Fan, Power Supply, and Disk LED States".


Pete

[cninham]

Thanks Pete. Will look at the doc.

Cameron

[Sridhar Bhaskarla]

I had this 'reseating" problem couple of times - particularly on Jamaica drives. However, there is not guarantee that it wouldn't show up again. Keep making make_tape_recoveries regularly because you don't know when it will go down. Also have a spare disk ready so incase of failure, you don't have to search for it.

-Sri

[Amiel Tutolo]

I just recently haad this same problem with my L2000. The led stayed a steady green. I could reseat and everything was fine. The CE for our area replaced the Media back plane without any success. I ended up just asking for a drive and then recreating the mirror back over from the alternate drive. Since oyu have the hotswap drives you can do this with the machine up and users on. I would suggest shutting down and replacing and then just boot off of your mirror. I have attached the HP doc that I used to recreate the mirror. Hope my one cents helps.