[Info-vax] Device with owner that doesn't exist.

[Richard Maher]

On 23/10/2019 8:53 pm, Fred. Zwarts wrote:
> Op 23.okt..2019 om 14:34 schreef Jan-Erik Söderholm:
>> Den 2019-10-23 kl. 14:16, skrev Simon Clubley:
>>> On 2019-10-23, Volker Halle <volker_halle at hotmail.com> wrote:
>>>> The OpenVMS Freeware V8 tool CLRREF may help as well:
>>>> 
>>>> https://www.digiater.nl/openvms/freeware/v80/clrref/
>>>> 
>>>> Use at your own risk.
>>>> 
>>> 
>>> On the plus side however, if it goes wrong, Jan-Erik will get
>>> the reboot he needs to clear the reference in a supported way.
>>> :-)
>>> 
>>> Seriously however, it's probably indeed best to make sure the
>>> system is idle and not in the middle of doing something important
>>> before running this tool just in case.
>>> 
>>> Simon.
>>> 
>> 
>> I'm looking at it right now and I'll try to force this hanging
>> device scenario on our test box, of course. But When it is really
>> needed, is when work places in the factory stands still and really
>> need to get going.
>> 
>> And it must be possible to use it in the COM files that the users
>> can submit from there menu system.
>> 
>> Today, I simply change the appliation startup COM file and assignes
>> another unused TNA device number and let the hanging one die by its
>> onw. Actually, the device in my first post above (TNA8883:) is gone
>> now, a couple of hours later. I changed the COM file to use
>> TNA8884: instead so that they get the printer online again and then
>> left it like that. And now TNA8883: is simply gone...
>> 
>> The TNA device points to an IP and port number of a label pinter
>> (Zebra). Sometimes when the printouts hangs, the users simply close
>> their session and that is where the user process goes away and the
>> TNA device gets hang, I guess.
>> 
>> I also have another idea where a pool of TNA numbers are used in a
>> round-robbin fashion. Then we can let any hanging devices just go
>> away by their own.
>> 
>> But I'll look at this first and we'll see...
>> 
>> 
> 
> Maybe the device was being deleted, but was waiting for an IO
> operation to complete. It might have been that after a TCP time-out,
> or when the printer disconnected, this IO completed and the device
> was deleted.
> 

Could try write/reading from it to free it up?