tz88 device status mounted dismount alloc foreign

[Werner Feurich]

I am unable to initialize a new tape on this drive. I get a reply that device is already mounted. A dismount/nounload device: gives me a reply that the device is not mounted. The device is also still showing the last volume label. The last backup failed and the cleaning light was on. I've used a cleaning tape which cleared the light. I don't have any processes out there using the drive.

Any ideas?

[Hoff]

OpenVMS version and platform?

Can you issue a SHOW DEVICE ddcu:/FULL and find out what the owning process id is?

If there is one. And if the process still exists.

I'm going to guess that this is a case of dangling ownership; of a tape device that is now allocated to a non-existent process (PID).

If it's a dangling process ownership, there are several approaches. First, look for a tool available from the support center that releases the drive ownership. Also look for relevant ECOs for the tape driver, for OpenVMS itself, and for MOUNT, as there have been kits released for this for various versions.

There can be another trigger for dangling ownership, and that's a process that has used SHARE privilege to override and access the device. Under normal operations, you'll want to have SHARE privilege disabled.

http://h71000.www7.hp.com/wizard/wiz_2785.html

Depending on your risk tolerance and your willingness to work with kernel-code (and a system crash and a reboot would technically also release the ownership), there's also:

http://groups.google.com/group/comp.os.vms/browse_thread/thread/4e40d42acaac0fe6

If your risk tolerance is lower, there can be a tool to release this dangling ownership acquired from the HP support center.

Or reboot.

And again, the above is a guess based on available information, and may or may not apply to this case.

Stephen Hoffman
HoffmanLabs

[atul sardana]

Dear Werner,

First you use
$>dismount device name only
if it shows again repeat message you check process id
$>show device name/full

which communicate this device and kill it by

$>stop/id=id
If it shows...
"%SYSTEM-W-NONEXPR, nonexistent process"
then you do some steps .....

(Is this device is internal or hot plugable or external?)
if it is hot plugable then you can plugout this Drive and plugin again and after
$>mc sysman io auto
$>show device name then try to use it
other wise it it is external then power OFF of it and ON again and try to use it

May be these helpfull for you...


Thanks n give points how you satisfied by this answer....

Atul sardana

[Andreas Vollmer]

Hi Werner,

From my point of view there could be several issues that cuases this problem.

1) The TZ88 can only read & write medias that are initialized with a TZ88 or lower such as TZ87, TZ86 etc.
Inserting a media initialized with a TZ89 or TZ90 will confuse the tape drive. It is unable to identify the BOT mark and spinning the media a lots of time in hope to find the BOT. It stops then by displaying the LED "Use Cleaning Tape". The inserted media is probably NOT defect.
Solution: remove or replace media - maybe use the cleaning tape.

2) The tape leader, it is the small piece of plastic the will pull-out the media dropped of the hook.
But normally all LED's are flashing if this happened.
==> I assume this is the less likely possibility of your problem.

3) As mentioned before the tape could be mounted of another cluster member.
In order to verify:
MCR SYSMAN SET ENV/CLU
DO SHOW DEVICE M????:
DO SHOW DEVICE /FULL M????:
EXIT
If one system is reporting the tape is mounted - you know what to do.
If not - it is getting tricky.
Therefore Hoff asked you for the OpenVMS Version and if you are running a cluster or not.
Be prepared for a reboot - as a last resort - the Microsoft approach - Sorry.
But lets wait for your feedback before we come to this stage.
Ciao
Andreas

[Werner Feurich]

Thank you all for the advice on this issue.

Here was our systems manager came to for a fix for right now.

the only steps needed to move the drive from mkc500 to mkc400 was to power down the drive, change the scsi id from 5 to 4 using the thumbwheel on the back and power the drive back on. then use the command 'mcr sysman io auto' to get the system to look out there and see the device in the new location.

So we're good for now. Thanks

[Werner Feurich]

the only steps needed to move the drive from mkc500 to mkc400 was to power down the drive, change the scsi id from 5 to 4 using the thumbwheel on the back and power the drive back on. then use the command 'mcr sysman io auto' to get the system to look out there and see the device in the new location.