Not able to mount or INIT a disk

[Joewee]

How to get this fixed?

$ mount/over=id drb0
%%%%%%%%%%% OPCOM 2-JUN-2009 14:20:46.78 %%%%%%%%%%%
Message from user SYSTEM on XXXX
device already allocated to another user

%MOUNT-I-OPRQST, device already allocated to another user
%%%%%%%%%%% OPCOM 2-JUN-2009 14:20:46.78 %%%%%%%%%%%
Request 5, from user SYSTEM on XXXX
device _$1$DRB0: (XXXX) is not available for mounting.

%MOUNT-I-OPRQST, device _$1$DRB0: (XXXX) is not available for mounting.
Interrupt

XXXX4$ stop
%MOUNT-I-OPRQSTCAN, operator request canceled
XXXX$
%%%%%%%%%%% OPCOM 2-JUN-2009 14:20:52.79 %%%%%%%%%%%
Request 5 was canceled

XXXX$ init $1$drb0: arc1
%SYSTEM-W-DEVALLOC, device already allocated to another user
XXXX$ sho dev drb0/full

Disk $1$DRB0: (XXXX), device type 2 Member RAID 0, is online, file-oriented
device, shareable, available to cluster, error logging is enabled.

Error count 0 Operations completed 1242
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W
Reference count 1 Default buffer size 512
Total blocks 35545088 Sectors per track 64
Total cylinders 34712 Tracks per cylinder 16
Allocation class 1
Device access control list:
(IDENTIFIER=PP$ARCH_READ,ACCESS=READ)

[Hakan Zanderau ( Anders]

Try,

$ MCR SYSMAN
SYSMAN> SET ENV/CLU
SYSMAN> DO SHOW DEV/FILE DRB0

This will show what process in the cluster has
an open file on the disk.

Hakan

[marsh_1]

hi,

unfortunately the only reliable method of fixing this is to reboot. ensure you are up to date with patches and firmware for this box and storage. do you know which process /user died while doing things with this device to leave it in this state ?

fwiw

[Hakan Zanderau ( Anders]

Oooops, maybe I was to fast reading again...

"available to cluster"

Is the node a clustermember ??

Hakan

[Joewee]

No Not in cluster... Standalone.

[Peter Zeiszler]

show dev drb0
show dev/files drb0

See if something has files open. If the files are installed in memory you will have to remove them from memory.

If a process that is active has the disk already mounted then check what that process is and contact the owner.

[Richard W Hunt]

A SHOW DEVICE/FILES $1$DRB0 might tell you what is open on the volume. Until you close what is open, you will never dismount that volume short of something drastic and dangerous.

What interests me is that your system somehow thinks this disk is in a cluster environment, yet you say it is not. Was it EVER in a cluster environment? Do you have VMSCLUSTER set to something other than 0 in SYSGEN for that node?

Is this by any chance a SAN device and some other member of the SAN currently has the device locked up in some way?

[Hoff]

Is this the OpenVMS Alpha V7.1-1H2 box that encountered severe disk data corruptions a while back? This one:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1343166

That OpenVMS Alpha box was badly corrupted.

And whether or not the AlphaServer 2000 4/275 and the SWXCR Mylex DAC960 hardware involved was failing was (and probably is) an open question here, too.

That OpenVMS Alpha V7.1-1H2 release (also) has various known errors that look rather similar to this one; disks could get locked in odd states, etc. There were a large number of ECO kits released for that version, and the V7.1-2 roll-up (and its respective ECO kits) then became available as part of simplifying the ECO management on V7.1, V7.1-1H1 and V7.1-1H2.

[RBrown_1]

Richard said "What interests me is that your system somehow thinks this disk is in a cluster environment".

My testing seems to indicate that even on a non-clustered system, show dev/full of most disk drives include the phrase "available to cluster".

FWIW

[Jur van der Burg]

This is the key:

>Reference count 1

The disk is not mounted, but there's a process with a reference to it. You can find out which one by doing this:

$ ana/sys
SDA> set out x.x
SDA> show proc/chan all
SDA> exit
$ search x.x drb0/window=20

Then you may find the offending process.

Jur.