I received two responses, both from HP, regarding the corrupt HW databases. The solution worked. Many thanks to:
Phillip Brown and Jason Orendorf.
Below is Phillip's response. Jason's was very similar.
Since you've got such a simple system, you can easily re-init the HW DBs. This will renumber all of your devices, so save that copy of your old hwmgr -show scsi output. You'll move new device special file names back to the old names after the initialization.
In single-user, after you issue the "mountroot", run:
# dn_setup -init
# dsfmgr -K
Note: If LSM is in use, then you'll need to recreate the mandatory LSM devices
you've just deleted with the init. Do that like this:
# mknod /dev/volconfig c 41 0
# mknod /dev/voltrace c 41 1
# mknod /dev/voliod c 41 2
# mknod /dev/volinfo c 41 3
The following should return no errors if this is successful.
# dsfmgr -v
You should be able to bcheckrc and come up normally. I'd reboot normally as a further check. If it fails, let me know. We'll supply you with a procedure for re-making your hardware databases from scratch.
Phil
-----Original Message-----
From: Hutcheson, Mike [mailto:Mike_Hutcheson_at_baylor.edu]
Sent: Thursday, June 19, 2003 3:51 PM
To: tru64-unix-managers_at_ornl.gov
Subject: Need help fixing possibly corrupt hardware database
Dear System Administrators,
Hi. In a period of crisis yesterday, I removed a TZ88 tape drive from a live standalone DS20E running Tru64 5.1A. It was the only device attached to a single-ended SCSI bus. The tape drive served as an emergency replacement for one that failed on a different system. Field service repaired the TZ88 that failed, I set the SCSI ID on the repaired drive to what the ID of the original unit was, connected the SCSI cable, and powered it on. When will I learn?
After loading a tape into the drive, I issued an mt seod to skip to end of data on the tape. After pressing return, the system promptly crashed. D'oh!
It then started rebooting. Right after the file systems mounted the following error messages appeared:
dsfmgr: ERROR: ADD inconsistant datum for registered device: kevm HWID-BNID was : 4-2, is: 72-15
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 4"
dsfmgr: ERROR: ADD inconstistant datum for registered device: kevm.pterm HWID-BNID was: 4-2, is: 72-15
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 4"
dsfmgr: ERROR: ADD inconstistant datum for registered device: kevm.debug HWID-BNID was: 4-2, is: 72-15
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 4"
dsfmgr: ERROR: ADD inconstistant datum for registered device: tty00 HWID-BNID was: 35-3, is: 102-16
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 35"
dsfmgr: ERROR: ADD inconstistant datum for registered device: tty01 HWID-BNID was: 37-4, is: 104-17
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 37"
dsfmgr: ERROR: ADD inconstistant datum for registered device: lp0 HWID-BNID was: 39-5, is: 106-18
NOTE: probable cause was previous failure of command: "dsfmgr -R hwid 39"
bcheckrc: Device Naming failed boot configure or verify.
Please correct the problem and continue or reboot
INIT: SINGLE-USER MODE
I can recreate the above errors with the following command:
dn_setup -boot
If I boot into single user mode, then type bcheckrc, I get the above errors. However, if I then type bcheckrc again a second time, it mounts the file systems without error messages. However, I can get dn_setup -boot to generate the message again.
Below is information I got from hwmgr and dsfmgr:
#hwmgr show scsi
SCSI DEVICE DEVICE DRIVER NUM DEVICE FIRST
HWID: DEVICEID HOSTNAME TYPE SUBTYPE OWNER PATH FILE VALID PATH
---------------------------------------------------------------------------------------------------------
53: 0 disk none 2 2 dsk0 [6/0/0]
54: 1 disk none 0 2 dsk1 [6/1/0]
55: 2 disk none 0 2 dsk2 [6/2/0]
56: 3 disk none 0 2 dsk3 [6/3/0]
122: 4 cdrom none 0 1 cdrom1 [1/0/0]
134: 5 tape none 0 1 tape2 [5/5/0]
#hwmgr view dev
SCSI DEVICE DEVICE DRIVER NUM DEVICE FIRST
HWID: DEVICEID HOSTNAME TYPE SUBTYPE OWNER PATH FILE VALID PATH
---------------------------------------------------------------------------------------------------------
53: 0 disk none 2 2 dsk0 [6/0/0]
54: 1 disk none 0 2 dsk1 [6/1/0]
55: 2 disk none 0 2 dsk2 [6/2/0]
56: 3 disk none 0 2 dsk3 [6/3/0]
122: 4 cdrom none 0 1 cdrom1 [1/0/0]
134: 5 tape none 0 1 tape2 [5/5/0]
#dsfmgr -vVF
dsfmgr -vVF gives the following output:
Secure Session Lock. At Thu Jun 19 15:58:51 2003
dsfmgr: verify with fix all datum for system (5.1A-0 1885) at /
Default File Tree:
OK.
Device Class Directory Default Database:
OK.
Device Category to Class Directory Database:
OK.
Dev directory structure:
OK.
Device Status Files:
OK.
Dev Nodes:
WARNING: status record not found for HWID 71: c 'scp_scsi'
NOTE: to fix, execute: "dsfmgr: -K"
WARNING: status record not found for HWID 72: c 'kevm'
WARNING: status record not found for HWID 72: c 'kevm.pterm'
WARNING: status record not found for HWID 72: c 'kevm.debug'
WARNING: status record not found for HWID 102: c 'tty##'
WARNING: status record not found for HWID 104: c 'tty##'
WARNING: status record not found for HWID 106: c 'lp##'
OK.
Command Status = 2 : No such file or directory
Release Session Lock at Thu Jun 19 15:58:51 2003
I tried dsfmgr -K mentioned above but it just gave me the dsmgr errors listed at the beginning of this email.
What does the problem appear to be and how may I go about fixing it?
Thanks very much for your help,
Mike Hutcheson (Mike_Hutcheson_at_baylor.edu)
Systems Manager (254) 710-4110
Baylor University - Electronic Library
Received on Thu Jun 19 2003 - 22:42:24 NZST