LVM error during boot

[Ilker Tarkan_1]

Hi,

My 785/C3000 running HPUX10.00 is causing me trouble during boot. More specifically it doesn't boot!

Here is what is happening;
- During a move, I took out the two disks from the machine. The boot disk was in scsi id 5, the other disk was in 6.

- I put the disks back, boot disk in 5, extra in 6. When I boot, it shows the following messages:

------------Start of message
Primary boot Path: FWSCSI.6.0
Alternate boot Path: SCSI.6.0

Process is booting the first available device
To discontinue, press any key within 10 seconds
10 seconds expired, Proceeding

Searching for device(s) with bootable media...

Attempting to boot Primary Path
Attempted boot failed, continuing search.

BOOTABLE DEVICE FOUND
FWSCSI.5.0

;
System HALTING during LVM configuration.
------------End of message

- Then I took the disks out, put them in again, but this time I put the boot disk in scsi id 6 and extra in scsi id 5.

- It booted successfuly. There were no problems accessing the disks. But I see an email sent to root. It says (compete text is in the attached file):

------------Start of message
Probable Cause / Recommended Action:

The device has been reset by a Bus Device Reset message, a hard reset condition, or a power-on reset.

Alternatively, a removable medium has been loaded or replaced.

Alternatively, the mode parameters, microcode, or inquiry data for the device have been changed.

Alternatively, the installed version of the device driver does not match that of the installed version of HP-UX. Install the correct version of the driver.
------------End of message

- Also with this configuration (boot disk in 6), machine freezes frequently. When it freezes I can remotely access it. In that case I try a reboot, but it never succeeds as it freezes during shutdown with the following message:

---------Start of message
LVM: vg[1]: pvnum=0 (dev_t=0x1f035000) is POWERFAILED
DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility has started receiving excessive errors from the I/O subsystem. .... use the logtool utility in stm to determine which I/O subsystem is logging excessive errors
---------End of message


I did some research and I see that people with similar problems usually had this after they removed/changed a component in the system. And it can be fixed by updating BDRA. But I don't want to do that. Any other suggestions?

Any help is appreciated.

Thanks,
Ilker Tarkan

[Steven E. Protter · Accepted solution]

The reason scsi id solved your problem was because your system was specifically configured to boot off of one of two boot devices. If the id's are wrong, it gets a bit confused.

You might want to check your boot configuration based on my last post in this thread:

lvlnboot -v

http://forums1.itrc.hp.com/service/forums/parseCurl.do?CURL=%2Fcm%2FQuestionAnswer%2F1%2C%2C0x14ba4c4958f64c4eb4827ca8efaaaa48%2C00.html&admit=716493758+1064423907261+28353475

The error in your log file is because you pulled the disks out and put them back in. This is informational and will go away when you reboot the system.

I assume these are hot-swappable drives.

SEP

[Helen French]

Changing SCSI ID of root disk will defenitely affect system booting. Since you are getting LVM and hardware error even after putting it back to original state, I would check and make sure that your hardware is okay. Check your termination and cables. Also check your disks with stm and dd tools. if you get any errors, get hardware support. I would be more concerned about the disk in vg01 - 0x1f035000 - device file: /dev/dsk/c3t5d0.

[Ilker Tarkan_1]

Thanks for the help SEP and Shiju!

I now understand that machine gets confused if the disks are changed. But I still don't understand why it does not boot in the original configuration (boot disk in scsi 5) and gives me the following message:

----start of message
Primary boot Path: FWSCSI.6.0
Alternate boot Path: SCSI.6.0
----end of message

Primary boot path should have been FWSCSI.5.0, right?

I tend to agree with Shiju that my disk may be damaged. Although I have no clue what it means, stm says my boot disk has 5 "Non-Medium Errors", the other has 0 errors. But then the question becomes how it can work when in a different scsi id if it is damaged?

By the way, both of the disks are inserted on a sled. There are no terminators or cables involved. Everything should be handled by the system.

Thanks,
Ilker Tarkan

[Helen French]

Again, If you change the SCSI ID of your root disk, it WILL NOT change the primary boot path which is actually stored on the system's NVRAM. These initial boot informations are not stored on your disk, but on the system board chips. So, if you want to change the primary boot path, change it in the boot console menu (by hitting any key while booting). You can view the current boot paths and can change it from there.

This can also be set from command prompt using 'setboot' command (man setboot(1m)). Once you set that to ID 5, the system will boot from the disk which has SCSI ID 5.

I would be concerned if the disk is giving error on stm or dd. Check it with a hardware support personnel. Also HP-UX version 10.00 is very very old.... you may need to consider installing 11.X